When available, the publication has been linked as PDF. To access, please click on the PDF symbol at the bottom right of each listed publication.
2020 |
|
| 1. | Serge Kas Hanna; Rawad Bitar; Parimal Parag; Venkat Dasari, ; Salim El, Rouayheb Adaptive Distributed Stochastic Gradient Descent for Minimizing Delay in the Presence of Stragglers Conference Forthcoming 45th International Conference on Acoustics, Speech, and Signal Processing, Barcelona, May 4-8, 2020, Forthcoming. Abstract | BibTeX | Tags: Electrical Communications Engineering | Links: @conference{@Parimal2020, title = {Adaptive Distributed Stochastic Gradient Descent for Minimizing Delay in the Presence of Stragglers}, author = {Serge Kas Hanna; Rawad Bitar; Parimal Parag; Venkat Dasari, and Salim El, Rouayheb}, url = {http://cps.iisc.ac.in/wp-content/uploads/2020/02/Adaptive-Distributed-Stochastic-Gradient-Descent-for-Minimizing-delay-in-the-presence-of-Stragglers.pdf}, year = {2020}, date = {2020-05-05}, booktitle = {45th International Conference on Acoustics, Speech, and Signal Processing, Barcelona, May 4-8, 2020}, abstract = {We consider the setting where a master wants to run a distributed stochastic gradient descent (SGD) algorithm on n workers each having a subset of the data. Distributed SGD may suffer from the effect of stragglers, i.e., slow or unresponsive workers who cause delays. One solution studied in the literature is to wait at each iteration for the responses of the fastest k < n workers before updating the model, where k is a fixed parameter. The choice of the value of k presents a trade-off between the runtime (i.e., convergence rate) of SGD and the error of the model. Towards optimizing the error-runtime trade-off, we investigate distributed SGD with adaptive k. We first design an adaptive policy for varying k that optimizes this trade-off based on an upper bound on the error as a function of the wallclock time which we derive. Then, we propose an algorithm for adaptive distributed SGD that is based on a statistical heuristic. We implement our algorithm and provide numerical simulations which confirm our intuition and theoretical analysis}, keywords = {Electrical Communications Engineering}, pubstate = {forthcoming}, tppubtype = {conference} } We consider the setting where a master wants to run a distributed stochastic gradient descent (SGD) algorithm on n workers each having a subset of the data. Distributed SGD may suffer from the effect of stragglers, i.e., slow or unresponsive workers who cause delays. One solution studied in the literature is to wait at each iteration for the responses of the fastest k < n workers before updating the model, where k is a fixed parameter. The choice of the value of k presents a trade-off between the runtime (i.e., convergence rate) of SGD and the error of the model. Towards optimizing the error-runtime trade-off, we investigate distributed SGD with adaptive k. We first design an adaptive policy for varying k that optimizes this trade-off based on an upper bound on the error as a function of the wallclock time which we derive. Then, we propose an algorithm for adaptive distributed SGD that is based on a statistical heuristic. We implement our algorithm and provide numerical simulations which confirm our intuition and theoretical analysis |
| 2. | Srikrishna Acharya; Amrutur Bharadwaj; Yogesh Simmhan; Aditya Gopalan; Parimal Parag; Himanshu, Tyagi CORNET: A Co-Simulation Middleware for Robot Networks Conference Forthcoming 12th International Conference on COMmunication Systems & NETworkS, Bengaluru, Jan 7-11, 2019, Forthcoming. Abstract | BibTeX | Tags: Robotics @conference{@Srikrishna20, title = {CORNET: A Co-Simulation Middleware for Robot Networks}, author = {Srikrishna Acharya; Amrutur Bharadwaj; Yogesh Simmhan; Aditya Gopalan; Parimal Parag; Himanshu,Tyagi}, year = {2020}, date = {2020-01-06}, booktitle = {12th International Conference on COMmunication Systems & NETworkS, Bengaluru, Jan 7-11, 2019}, abstract = {This paper describes CORNET, a co-simulation middleware for applications involving multi-robot systems like a network of Unmanned Aerial Vehicle (UAV) systems. Design of such systems requires knowledge of the flight dynamics of UAVs and the communication links connecting UAVs with each other or with the ground control station. Besides, UAV networks are dynamic and distinctive from other ad-hoc networks and require protocols that can adapt to high-mobility, dynamic topology and changing link quality in power constrained resource platforms. Therefore, it is necessary to co-design the UAV path planning algorithms and the communication protocols. The proposed co-simulation framework integrates existing tools to simulate flight dynamics and network related aspects. Gazebo with robot operating system (ROS) is used as a physical system UAV simulator and NS-3 is used as a network simulator, to jointly capture the cyber-physical system (CPS) aspects of the multi- UAV systems. A particular aspect we address is on synchronizing time and position across the two simulation environments, and we provide APIs to allow easy migration of the algorithms to real platforms.}, keywords = {Robotics}, pubstate = {forthcoming}, tppubtype = {conference} } This paper describes CORNET, a co-simulation middleware for applications involving multi-robot systems like a network of Unmanned Aerial Vehicle (UAV) systems. Design of such systems requires knowledge of the flight dynamics of UAVs and the communication links connecting UAVs with each other or with the ground control station. Besides, UAV networks are dynamic and distinctive from other ad-hoc networks and require protocols that can adapt to high-mobility, dynamic topology and changing link quality in power constrained resource platforms. Therefore, it is necessary to co-design the UAV path planning algorithms and the communication protocols. The proposed co-simulation framework integrates existing tools to simulate flight dynamics and network related aspects. Gazebo with robot operating system (ROS) is used as a physical system UAV simulator and NS-3 is used as a network simulator, to jointly capture the cyber-physical system (CPS) aspects of the multi- UAV systems. A particular aspect we address is on synchronizing time and position across the two simulation environments, and we provide APIs to allow easy migration of the algorithms to real platforms. |
| 3. | Nihesh Rathod; Renu Subramanian; Rajesh, Sundaresan Coverage Estimation in Outdoor Heterogeneous Propagation Environments Journal Article IEEE Access, 8 , pp. 31660 - 31673, 2020. Abstract | BibTeX | Tags: Electrical Communications Engineering | Links: @article{@Nihesh2020, title = {Coverage Estimation in Outdoor Heterogeneous Propagation Environments}, author = {Nihesh Rathod; Renu Subramanian; Rajesh,Sundaresan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2020/03/Coverage-Estimation-in-Outdoor-Heterogeneous-Propagation-Environments.pdf}, doi = { 10.1109/ACCESS.2020.2972811}, year = {2020}, date = {2020-02-10}, journal = { IEEE Access}, volume = {8}, pages = { 31660 - 31673}, abstract = {This paper is on a coverage estimation procedure for the deployment of outdoor Internet of Things (IoT). In the first part of the paper, a data-driven coverage estimation technique is proposed. The estimation technique combines multiple machine-learning-based regression ideas. The proposed technique achieves two purposes. The first purpose is to reduce the bias in the estimated received signal strength arising from estimations performed only on the successfully received packets. The second purpose is to exploit commonality of physical parameters, e.g. antenna-gain, in measurements that are made across multiple propagation environments. It also provides the correct link function for performing a nonlinear regression in our communication systems context. In the second part of the paper, a method to use readily available geographic information system (GIS) data (for classifying geographic areas into various propagation environments) followed by an algorithm for estimating received signal strength (which is motivated by the first part of the paper) is proposed. Together they enable quick and automated estimation of coverage in outdoor environments. It is anticipated that these will lead to faster and more efficient deployment of outdoor. Internet of Things}, keywords = {Electrical Communications Engineering}, pubstate = {published}, tppubtype = {article} } This paper is on a coverage estimation procedure for the deployment of outdoor Internet of Things (IoT). In the first part of the paper, a data-driven coverage estimation technique is proposed. The estimation technique combines multiple machine-learning-based regression ideas. The proposed technique achieves two purposes. The first purpose is to reduce the bias in the estimated received signal strength arising from estimations performed only on the successfully received packets. The second purpose is to exploit commonality of physical parameters, e.g. antenna-gain, in measurements that are made across multiple propagation environments. It also provides the correct link function for performing a nonlinear regression in our communication systems context. In the second part of the paper, a method to use readily available geographic information system (GIS) data (for classifying geographic areas into various propagation environments) followed by an algorithm for estimating received signal strength (which is motivated by the first part of the paper) is proposed. Together they enable quick and automated estimation of coverage in outdoor environments. It is anticipated that these will lead to faster and more efficient deployment of outdoor. Internet of Things |
| 4. | Kolathaya, Shishir Local stability of PD controlled bipedal walking robots Journal Article Automatica, 114 (108841), 2020. Abstract | BibTeX | Tags: Robotics | Links: @article{@Shishir2020, title = {Local stability of PD controlled bipedal walking robots}, author = {Shishir Kolathaya}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/12/PD-based-Robust-Quadratic-Programs-for-Robotic-Systems.pdf}, doi = {https://doi.org/10.1016/j.automatica.2020.108841}, year = {2020}, date = {2020-04-01}, journal = {Automatica}, volume = {114}, number = {108841}, abstract = {We establish stability results for PD tracking control laws in bipedal walking robots. Stability of PD control laws for continuous robotic systems is an established result, and we extend this for hybrid robotic systems, an alternating sequence of continuous and discrete events. Bipedal robots have the leg-swing as the continuous event, and the foot-strike as the discrete event. In addition, bipeds largely have underactuations due to the interactions between feet and ground. For each continuous event, we establish that the convergence rate of the tracking error can be regulated via appropriate tuning of the PD gains; and for each discrete event, we establish that this convergence rate sufficiently overcomes the nonlinear impacts by assumptions on the hybrid zero dynamics. The main contributions are (1) Extension of the stability results of PD control laws for underactuated robotic systems, and (2) Exponential ultimate boundedness of hybrid periodic orbits under the assumption of exponential stability of their projections to the hybrid zero dynamics. Towards the end, we will validate these results in a 2-link bipedal walker in simulation.}, keywords = {Robotics}, pubstate = {published}, tppubtype = {article} } We establish stability results for PD tracking control laws in bipedal walking robots. Stability of PD control laws for continuous robotic systems is an established result, and we extend this for hybrid robotic systems, an alternating sequence of continuous and discrete events. Bipedal robots have the leg-swing as the continuous event, and the foot-strike as the discrete event. In addition, bipeds largely have underactuations due to the interactions between feet and ground. For each continuous event, we establish that the convergence rate of the tracking error can be regulated via appropriate tuning of the PD gains; and for each discrete event, we establish that this convergence rate sufficiently overcomes the nonlinear impacts by assumptions on the hybrid zero dynamics. The main contributions are (1) Extension of the stability results of PD control laws for underactuated robotic systems, and (2) Exponential ultimate boundedness of hybrid periodic orbits under the assumption of exponential stability of their projections to the hybrid zero dynamics. Towards the end, we will validate these results in a 2-link bipedal walker in simulation. |
| 5. | Pratik Sharma; Devam Awasare; Bishal Jaiswal; Srivats Mohan; Abinaya N; Ishan Darwhekar; Anand SVR; Bharadwaj Amrutur; Aditya Gopalan; Parimal Parag; Himanshu, Tyagi On the Latency in Vehicular Control using Video Streaming over Wi-Fi Conference Forthcoming 26th National Conference on Communications, IIT Kharagpur, 21-23 February 2020, Forthcoming. Abstract | BibTeX | Tags: Electrical Communications Engineering | Links: @conference{@Pratik2020, title = {On the Latency in Vehicular Control using Video Streaming over Wi-Fi}, author = {Pratik Sharma; Devam Awasare; Bishal Jaiswal; Srivats Mohan; Abinaya,N; Ishan Darwhekar; Anand SVR; Bharadwaj Amrutur; Aditya Gopalan; Parimal Parag; Himanshu,Tyagi}, url = {http://cps.iisc.ac.in/wp-content/uploads/2020/02/On-the-Latency-in-Vehicular-Control-using-Video-Streaming-over-Wi-Fi.pdf}, year = {2020}, date = {2020-02-23}, booktitle = {26th National Conference on Communications, IIT Kharagpur, 21-23 February 2020}, abstract = {We consider the use of Wi-Fi (IEEE 802.11n/r) network for remote control of a vehicle using video transmission on the uplink and control signals for the actuator on the downlink. We have setup a network with multiple access points (AP) providing indoor and outdoor coverage, which connects an unmanned ground vehicle (UGV) to a remote command center. Additionally, our setup includes a redundant IEEE 802.11p link for sending control messages over downlink with high reliability and low latency. We study the end-to-end communication delay and complete a latency profiling for each sub-component, including the video codec and the Wi-Fi links. Furthermore, we provide guidelines for practical design choices including the optimal configuration of the scanning process during handoffs and the codec parameters for delay optimization. Overall, our proposed configuration reduces the end-to-end delay significantly in comparison with the default configuration.}, keywords = {Electrical Communications Engineering}, pubstate = {forthcoming}, tppubtype = {conference} } We consider the use of Wi-Fi (IEEE 802.11n/r) network for remote control of a vehicle using video transmission on the uplink and control signals for the actuator on the downlink. We have setup a network with multiple access points (AP) providing indoor and outdoor coverage, which connects an unmanned ground vehicle (UGV) to a remote command center. Additionally, our setup includes a redundant IEEE 802.11p link for sending control messages over downlink with high reliability and low latency. We study the end-to-end communication delay and complete a latency profiling for each sub-component, including the video codec and the Wi-Fi links. Furthermore, we provide guidelines for practical design choices including the optimal configuration of the scanning process during handoffs and the codec parameters for delay optimization. Overall, our proposed configuration reduces the end-to-end delay significantly in comparison with the default configuration. |
2019 |
|
| 6. | Chakraborty, Atlanta; Chandru, Vijay; Rao, M R A linear programming primer: From Fourier to Karmarkar Journal Article Annals of Operations Research, 2019. Abstract | BibTeX | Tags: Control and Optimisation | Links: @article{Chakraborty2019, title = {A linear programming primer: From Fourier to Karmarkar}, author = {Atlanta Chakraborty and Vijay Chandru and M. R. Rao}, url = {http://www.rbccps.org/wp-content/uploads/2019/03/Chakraborty2019_Article_ALinearProgrammingPrimerFromFo.pdf}, doi = {10.1007/s10479-019-03186-2}, year = {2019}, date = {2019-03-01}, journal = {Annals of Operations Research}, abstract = {The story of linear programming is one with all the elements of a grand historical drama. The original idea of testing if a polyhedron is non-empty by using a variable elimination to project down one dimension at a time until a tautology emerges dates back to a paper by Fourier in 1823. This gets re-invented in the 1930s by Motzkin. The real interest in linear programming happens during World War II when mathematicians ponder best ways of utilising resources at a time when they are constrained. The problem of optimising a linear function over a set of linear inequalities becomes the focus of the effort. Dantzig’s Simplex Method is announced and the Rand Corporation becomes a hot bed of computational mathematics. The range of applications of this modelling approach grows and the powerful machinery of numerical analysis and numerical linear algebra becomes a major driver for the advancement of computing machines. In the 1970s, constructs of theoretical computer science indicate that linear programming may in fact define the frontier of tractable problems that can be solved effectively on large instances. This raised a series of questions and answers: Is the Simplex Method a polynomial-time method and if not can we construct novel polynomial time methods, etc. And that is how the Ellipsoid Method from the Soviet Union and the Interior Point Method from Bell Labs make their way into this story as the heroics of Khachiyan and Karmarkar. We have called this paper a primer on linear programming since it only gives the reader a quick narrative of the grand historical drama. Hopefully it motivates a young reader to delve deeper and add another chapter.}, keywords = {Control and Optimisation}, pubstate = {published}, tppubtype = {article} } The story of linear programming is one with all the elements of a grand historical drama. The original idea of testing if a polyhedron is non-empty by using a variable elimination to project down one dimension at a time until a tautology emerges dates back to a paper by Fourier in 1823. This gets re-invented in the 1930s by Motzkin. The real interest in linear programming happens during World War II when mathematicians ponder best ways of utilising resources at a time when they are constrained. The problem of optimising a linear function over a set of linear inequalities becomes the focus of the effort. Dantzig’s Simplex Method is announced and the Rand Corporation becomes a hot bed of computational mathematics. The range of applications of this modelling approach grows and the powerful machinery of numerical analysis and numerical linear algebra becomes a major driver for the advancement of computing machines. In the 1970s, constructs of theoretical computer science indicate that linear programming may in fact define the frontier of tractable problems that can be solved effectively on large instances. This raised a series of questions and answers: Is the Simplex Method a polynomial-time method and if not can we construct novel polynomial time methods, etc. And that is how the Ellipsoid Method from the Soviet Union and the Interior Point Method from Bell Labs make their way into this story as the heroics of Khachiyan and Karmarkar. We have called this paper a primer on linear programming since it only gives the reader a quick narrative of the grand historical drama. Hopefully it motivates a young reader to delve deeper and add another chapter. |
| 7. | P T, Akhil; Eitan Altman; Rajesh Sundaresan A Mean-field Approach for Controlling Singularly Perturbed Multi-population SIS Epidemics Journal Article arXiv.org: Math, 2019. Abstract | BibTeX | Tags: Control and Optimisation | Links: @article{Akhil2019, title = {A Mean-field Approach for Controlling Singularly Perturbed Multi-population SIS Epidemics}, author = {P T, Akhil; Eitan Altman; Rajesh Sundaresan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/A-mean-fields-approach-for-controlling-singularly-perturbed-multi-population-SIS-Epidemics.pdf}, year = {2019}, date = {2019-02-15}, journal = {arXiv.org: Math}, abstract = {We consider a multi-population epidemic model with one or more (almost) isolated communities and one mobile community. Each of the isolated communities has contact within itself and, in addition, contact with the outside world but only through the mobile community. The contact rate between the mobile community and the other communities is assumed to be controlled. We first derive a multidimensional ordinary differential equation (ODE) as a mean-field fluid approximation to the process of the number of infected nodes, after appropriate scaling. We show that the approximation becomes tight as the sizes of the communities grow. We then use a singular perturbation approach to reduce the dimension of the ODE and identify an optimal control policy on this system over a fixed time horizon via Pontryagin’s minimum principle. We then show that this policy is close to optimal, within a certain class, on the original problem for large enough communities. From a phenomenological perspective, we show that the epidemic may sustain in time in all communities (and thus the system has a nontrivial metastable regime) even though in the absence of the mobile nodes the epidemic would die out quickly within each of the isolated communities.}, keywords = {Control and Optimisation}, pubstate = {published}, tppubtype = {article} } We consider a multi-population epidemic model with one or more (almost) isolated communities and one mobile community. Each of the isolated communities has contact within itself and, in addition, contact with the outside world but only through the mobile community. The contact rate between the mobile community and the other communities is assumed to be controlled. We first derive a multidimensional ordinary differential equation (ODE) as a mean-field fluid approximation to the process of the number of infected nodes, after appropriate scaling. We show that the approximation becomes tight as the sizes of the communities grow. We then use a singular perturbation approach to reduce the dimension of the ODE and identify an optimal control policy on this system over a fixed time horizon via Pontryagin’s minimum principle. We then show that this policy is close to optimal, within a certain class, on the original problem for large enough communities. From a phenomenological perspective, we show that the epidemic may sustain in time in all communities (and thus the system has a nontrivial metastable regime) even though in the absence of the mobile nodes the epidemic would die out quickly within each of the isolated communities. |
| 8. | Kuraganti Chetan Kumar; Gurunath Gurrala; Rajesh, Sundaresan A Sequential Testing Framework for Identifying a Transmission Line Outage in a Power System Conference e-Energy '19 Proceedings of the Tenth ACM International Conference on Future Energy Systems, June 25 - 28, 2019, Phoenix, AZ, USA, 2019. Abstract | BibTeX | Tags: Smart Grid | Links: @conference{@Chetan2019, title = {A Sequential Testing Framework for Identifying a Transmission Line Outage in a Power System}, author = {Kuraganti Chetan Kumar; Gurunath Gurrala; Rajesh,Sundaresan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/A-Sequential-Testing-Framework-for-Identifying-a-Transmission-Line-Outage-in-a-Power-System.pdf}, doi = {10.1145/3307772.3328297}, year = {2019}, date = {2019-06-15}, booktitle = {e-Energy '19 Proceedings of the Tenth ACM International Conference on Future Energy Systems, June 25 - 28, 2019, Phoenix, AZ, USA}, pages = {331-342}, abstract = {The topology of a power system changes when a line outage is encountered. Identifying which line has failed in the shortest possible time is of importance due to the cascading nature of such failures. In this work, we propose a state estimation based sequential hypothesis testing procedure to locate the failed line. We focus on single line outages as these are the most frequently occurring failures. Earlier work on state estimation based sequential testing procedure used a DC approximation model assuming that the sensors provided angle and voltage information. This is known to be a coarse model but results in a simpler linear estimation problem. In this work, we look at a finer nonlinear model of power measurements and treat phase angles and voltages as hidden states. After a local linearization, we propose a Kalman filter based state estimation followed by a generalized likelihood ratio testing procedure to determine which of the lines has failed. We consider both centralized and decentralized approaches. In the centralized case, measurements from every installed meter is made available to the system operator. In the decentralized case, only limited aggregated information is made available because of, for example, communication capacity constraints. We test our algorithms on the IEEE 14 and 118 bus systems and show that all high risk link failures are quickly identified.}, keywords = {Smart Grid}, pubstate = {published}, tppubtype = {conference} } The topology of a power system changes when a line outage is encountered. Identifying which line has failed in the shortest possible time is of importance due to the cascading nature of such failures. In this work, we propose a state estimation based sequential hypothesis testing procedure to locate the failed line. We focus on single line outages as these are the most frequently occurring failures. Earlier work on state estimation based sequential testing procedure used a DC approximation model assuming that the sensors provided angle and voltage information. This is known to be a coarse model but results in a simpler linear estimation problem. In this work, we look at a finer nonlinear model of power measurements and treat phase angles and voltages as hidden states. After a local linearization, we propose a Kalman filter based state estimation followed by a generalized likelihood ratio testing procedure to determine which of the lines has failed. We consider both centralized and decentralized approaches. In the centralized case, measurements from every installed meter is made available to the system operator. In the decentralized case, only limited aggregated information is made available because of, for example, communication capacity constraints. We test our algorithms on the IEEE 14 and 118 bus systems and show that all high risk link failures are quickly identified. |
| 9. | Karanjkar, Neha; Tejasvi, Poorna Chandra; Amrutur, Bharadwaj A simpy-based simulation testbed Presentation 18.04.2019. Abstract | BibTeX | Tags: Smart City | Links: @misc{Karanjkar2019, title = {A simpy-based simulation testbed }, author = {Neha Karanjkar and Poorna Chandra Tejasvi and Bharadwaj Amrutur}, doi = {10.1145/3302505.3312591}, year = {2019}, date = {2019-04-18}, abstract = {A real-time testbed that emulates a large number of IoT end-points generating traffic to the middleware/application layers can be used for debugging and performance evaluation of the smart-city software platforms prior to deployment. We propose an architecture for such a simulation testbed based on Python's SimPy library. The simulated IoT end-points communicate with the middleware in real-time and can also interact directly with each other and with a common shared environment. This makes the testbed particularly suited for modeling system-wide scenarios such as synchronized faults and power outages.}, keywords = {Smart City}, pubstate = {published}, tppubtype = {presentation} } A real-time testbed that emulates a large number of IoT end-points generating traffic to the middleware/application layers can be used for debugging and performance evaluation of the smart-city software platforms prior to deployment. We propose an architecture for such a simulation testbed based on Python's SimPy library. The simulated IoT end-points communicate with the middleware in real-time and can also interact directly with each other and with a common shared environment. This makes the testbed particularly suited for modeling system-wide scenarios such as synchronized faults and power outages. |
| 10. | Raghuram Bharadwaj Diddigi ; D Sai Koti Reddy ; K J, Prabhuchandran ;Shalabh Bhatnagar Actor-Critic Algorithms for Constrained Multi-agent Reinforcement Learning Conference AAMAS 19, Proceedings of the 18th International Conference on Autonomous Agents and MultiAgent Systems, 2019, ISBN: 978-1-4503-6309-9. Abstract | BibTeX | Tags: Multiagent Systems | Links: @conference{Raghuram2019, title = {Actor-Critic Algorithms for Constrained Multi-agent Reinforcement Learning}, author = {Raghuram Bharadwaj Diddigi ; D, Sai Koti Reddy ; K J, Prabhuchandran ;Shalabh Bhatnagar}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/Actor-Critic-Algorithms-for-Constrained-multi-agent-Reinforcement-Learning.pdf}, isbn = {978-1-4503-6309-9}, year = {2019}, date = {2019-05-08}, booktitle = {AAMAS 19, Proceedings of the 18th International Conference on Autonomous Agents and MultiAgent Systems}, journal = {AAMAS (International Conference on Autonomous Agents and Multiagent Systems)}, pages = {1931-1933 }, abstract = {Multi-agent reinforcement learning has gained lot of popularity primarily owing to the success of deep function approximation architectures. However, many real-life multi-agent applications often impose constraints on the joint action sequence that can be taken by the agents. In this work, we formulate such problems in the framework of constrained cooperative stochastic games. Under this setting, the goal of the agents is to obtain joint action sequence that minimizes a total cost objective criterion subject to total cost penalty/budget functional constraints. To this end, we utilize the Lagrangian formulation and propose actor-critic algorithms. Through experiments on a constrained multi-agent grid world task, we demonstrate that our algorithms converge to near-optimal joint action sequences satisfying the given constraints.}, keywords = {Multiagent Systems}, pubstate = {published}, tppubtype = {conference} } Multi-agent reinforcement learning has gained lot of popularity primarily owing to the success of deep function approximation architectures. However, many real-life multi-agent applications often impose constraints on the joint action sequence that can be taken by the agents. In this work, we formulate such problems in the framework of constrained cooperative stochastic games. Under this setting, the goal of the agents is to obtain joint action sequence that minimizes a total cost objective criterion subject to total cost penalty/budget functional constraints. To this end, we utilize the Lagrangian formulation and propose actor-critic algorithms. Through experiments on a constrained multi-agent grid world task, we demonstrate that our algorithms converge to near-optimal joint action sequences satisfying the given constraints. |
| 11. | Jogendra Nath Kundu; Maharshi Gor; R, Venkatesh Babu BiHMP-GAN: Bidirectional 3D Human Motion Prediction GAN Conference Proceedings of the Thirty-Third AAAI Conference on Artificial Intelligence, 33 , 2019. Abstract | BibTeX | Tags: Student Research Grant | Links: @conference{Jogendra2019, title = {BiHMP-GAN: Bidirectional 3D Human Motion Prediction GAN}, author = {Jogendra Nath Kundu; Maharshi Gor; R, Venkatesh Babu}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/BiHMP-GAN-Bidirectional-3D-human-motion-Prediction-GAN.pdf}, doi = {10.1609/aaai.v33i01.33018553}, year = {2019}, date = {2019-07-17}, booktitle = {Proceedings of the Thirty-Third AAAI Conference on Artificial Intelligence}, volume = {33}, abstract = {Human motion prediction model has applications in various fields of computer vision. Without taking into account the inherent stochasticity in the prediction of future pose dynamics, such methods often converges to a deterministic undesired mean of multiple probable outcomes. Devoid of this, we propose a novel probabilistic generative approach called Bidirectional Human motion prediction GAN, or BiHMP-GAN. To be able to generate multiple probable human-pose sequences, conditioned on a given starting sequence, we introduce a random extrinsic factor r, drawn from a predefined prior distribution. Furthermore, to enforce a direct content loss on the predicted motion sequence and also to avoid mode-collapse, a novel bidirectional framework is incorporated by modifying the usual discriminator architecture. The discriminator is trained also to regress this extrinsic factor r, which is used alongside with the intrinsic factor (encoded starting pose sequence) to generate a particular pose sequence. To further regularize the training, we introduce a novel recursive prediction strategy. In spite of being in a probabilistic framework, the enhanced discriminator architecture allows predictions of an intermediate part of pose sequence to be used as a conditioning for prediction of the latter part of the sequence. The bidirectional setup also provides a new direction to evaluate the prediction quality against a given test sequence. For a fair assessment of BiHMP-GAN, we report performance of the generated motion sequence using (i) a critic model trained to discriminate between real and fake motion sequence, and (ii) an action classifier trained on real human motion dynamics. Outcomes of both qualitative and quantitative evaluations, on the probabilistic generations of the model, demonstrate the superiority of BiHMP-GAN over previously available methods.}, keywords = {Student Research Grant}, pubstate = {published}, tppubtype = {conference} } Human motion prediction model has applications in various fields of computer vision. Without taking into account the inherent stochasticity in the prediction of future pose dynamics, such methods often converges to a deterministic undesired mean of multiple probable outcomes. Devoid of this, we propose a novel probabilistic generative approach called Bidirectional Human motion prediction GAN, or BiHMP-GAN. To be able to generate multiple probable human-pose sequences, conditioned on a given starting sequence, we introduce a random extrinsic factor r, drawn from a predefined prior distribution. Furthermore, to enforce a direct content loss on the predicted motion sequence and also to avoid mode-collapse, a novel bidirectional framework is incorporated by modifying the usual discriminator architecture. The discriminator is trained also to regress this extrinsic factor r, which is used alongside with the intrinsic factor (encoded starting pose sequence) to generate a particular pose sequence. To further regularize the training, we introduce a novel recursive prediction strategy. In spite of being in a probabilistic framework, the enhanced discriminator architecture allows predictions of an intermediate part of pose sequence to be used as a conditioning for prediction of the latter part of the sequence. The bidirectional setup also provides a new direction to evaluate the prediction quality against a given test sequence. For a fair assessment of BiHMP-GAN, we report performance of the generated motion sequence using (i) a critic model trained to discriminate between real and fake motion sequence, and (ii) an action classifier trained on real human motion dynamics. Outcomes of both qualitative and quantitative evaluations, on the probabilistic generations of the model, demonstrate the superiority of BiHMP-GAN over previously available methods. |
| 12. | K L Navaneet; Priyanka Mandikal; Mayank Agarwal; R, Venkatesh Babu CAPNet: Continuous Approximation Projection for 3D Point Cloud Reconstruction Using 2D Supervision Conference Proceedings of the AAAI Conference on Artificial Intelligence, 33 , 2019. Abstract | BibTeX | Tags: Student Research Grant | Links: @conference{Navaneet2019, title = {CAPNet: Continuous Approximation Projection for 3D Point Cloud Reconstruction Using 2D Supervision}, author = {K L, Navaneet; Priyanka Mandikal; Mayank Agarwal; R, Venkatesh Babu}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/Capnet-Continuous-Approximation-projection-for-3D-point-Cloud-Reconstruction-using-2D-Supervision.pdf}, doi = {10.1609/aaai.v33i01.33018819}, year = {2019}, date = {2019-07-17}, booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence}, volume = {33}, pages = {8819-8826}, abstract = {Knowledge of 3D properties of objects is a necessity in order to build effective computer vision systems. However, lack of large scale 3D datasets can be a major constraint for datadriven approaches in learning such properties. We consider the task of single image 3D point cloud reconstruction, and aim to utilize multiple foreground masks as our supervisory data to alleviate the need for large scale 3D datasets. A novel differentiable projection module, called ‘CAPNet’, is introduced to obtain such 2D masks from a predicted 3D point cloud. The key idea is to model the projections as a continuous approximation of the points in the point cloud. To overcome the challenges of sparse projection maps, we propose a loss formulation termed ‘affinity loss’ to generate outlierfree reconstructions. We significantly outperform the existing projection based approaches on a large-scale synthetic dataset. We show the utility and generalizability of such a 2D supervised approach through experiments on a real-world dataset, where lack of 3D data can be a serious concern. To further enhance the reconstructions, we also propose a test stage optimization procedure to obtain reconstructions that display high correspondence with the observed input image.}, keywords = {Student Research Grant}, pubstate = {published}, tppubtype = {conference} } Knowledge of 3D properties of objects is a necessity in order to build effective computer vision systems. However, lack of large scale 3D datasets can be a major constraint for datadriven approaches in learning such properties. We consider the task of single image 3D point cloud reconstruction, and aim to utilize multiple foreground masks as our supervisory data to alleviate the need for large scale 3D datasets. A novel differentiable projection module, called ‘CAPNet’, is introduced to obtain such 2D masks from a predicted 3D point cloud. The key idea is to model the projections as a continuous approximation of the points in the point cloud. To overcome the challenges of sparse projection maps, we propose a loss formulation termed ‘affinity loss’ to generate outlierfree reconstructions. We significantly outperform the existing projection based approaches on a large-scale synthetic dataset. We show the utility and generalizability of such a 2D supervised approach through experiments on a real-world dataset, where lack of 3D data can be a serious concern. To further enhance the reconstructions, we also propose a test stage optimization procedure to obtain reconstructions that display high correspondence with the observed input image. |
| 13. | Gayarthi, H; Gulhare, Siddhartha; Verma, Ashish Characteristics of stop and go wave in one dimensional interrupted pedestrian flow through narrow channel Journal Article Collective Dynamics, pp. A18:1-14, 2019. Abstract | BibTeX | Tags: The Kumbh Mela Experiment | Links: @article{Gayarthi2019, title = {Characteristics of stop and go wave in one dimensional interrupted pedestrian flow through narrow channel}, author = {H. Gayarthi and Siddhartha Gulhare and Ashish Verma}, url = {http://www.rbccps.org/wp-content/uploads/2019/05/34-296-1-PB.pdf}, doi = {10.17815/CD.2018.18}, year = {2019}, date = {2019-01-12}, journal = {Collective Dynamics}, pages = {A18:1-14}, abstract = {Pedestrian microscopic simulation models can aid crowd management only if they can reproduce the crowd behavior correctly. To calibrate and validate the model, it is important to understand crowd movement during various activities involved in mass gathering events. A common practice in such gathering is to hold attendees in waiting area in near corridors separated by crowd barriers before the event and allow entering the event only after a designated time. The crowd is released in small batches to avoid overcrowding inside. Long waiting hours, anger, excitement, competitive feeling etc. can make crowd aggressive during such entries. Crowd flow characteristics due to such behavior is difficult to recreate in pedestrian experimental studies in laboratory setting. This paper studied interrupted flow of such crowd through a narrow corridors made of strong railing channel inside a temple. Interrupted flow lead to formation of one dimensional stop and go waves. These stop and go waves were studied from the trajectory data. The average speed of waves propagating over longer distance were also estimated. The quantitative output from this study can be used to calibrate and validate simulation models of such activity during mass gathering events.}, keywords = {The Kumbh Mela Experiment}, pubstate = {published}, tppubtype = {article} } Pedestrian microscopic simulation models can aid crowd management only if they can reproduce the crowd behavior correctly. To calibrate and validate the model, it is important to understand crowd movement during various activities involved in mass gathering events. A common practice in such gathering is to hold attendees in waiting area in near corridors separated by crowd barriers before the event and allow entering the event only after a designated time. The crowd is released in small batches to avoid overcrowding inside. Long waiting hours, anger, excitement, competitive feeling etc. can make crowd aggressive during such entries. Crowd flow characteristics due to such behavior is difficult to recreate in pedestrian experimental studies in laboratory setting. This paper studied interrupted flow of such crowd through a narrow corridors made of strong railing channel inside a temple. Interrupted flow lead to formation of one dimensional stop and go waves. These stop and go waves were studied from the trajectory data. The average speed of waves propagating over longer distance were also estimated. The quantitative output from this study can be used to calibrate and validate simulation models of such activity during mass gathering events. |
| 14. | Dholakiya, Dhaivat; Bhattacharya, Shounak; Gunalan, Ajay; Singla, Abhik; Bhatnagar, Shalabh; Amrutur, Bharadwaj; Ghosal, Ashitava; Kolathaya, Shishir Design, development and experimental realization of a quadrupedal research platform: Stoch Journal Article arXiv: Computer Science, 2019. Abstract | BibTeX | Tags: Walking Robot | Links: @article{Dholakiya2019, title = {Design, development and experimental realization of a quadrupedal research platform: Stoch}, author = {Dhaivat Dholakiya and Shounak Bhattacharya and Ajay Gunalan and Abhik Singla and Shalabh Bhatnagar and Bharadwaj Amrutur and Ashitava Ghosal and Shishir Kolathaya}, url = {http://www.rbccps.org/wp-content/uploads/2019/05/1901.00697.pdf}, year = {2019}, date = {2019-01-03}, journal = {arXiv: Computer Science}, abstract = {In this paper, we present a complete description of the hardware design and control architecture of our custom built quadruped robot, called the `Stoch'. Our goal is to realize a robust, modular, and a reliable quadrupedal platform, using which various locomotion behaviors are explored. This platform enables us to explore different research problems in legged locomotion, which use both traditional and learning based techniques. We discuss the merits and limitations of the platform in terms of exploitation of available behaviours, fast rapid prototyping, reproduction and repair. Towards the end, we will demonstrate trotting, bounding behaviors, and preliminary results in turning. In addition, we will also show various gait transitions i.e., trot-to-turn and trot-to-bound behaviors. }, keywords = {Walking Robot}, pubstate = {published}, tppubtype = {article} } In this paper, we present a complete description of the hardware design and control architecture of our custom built quadruped robot, called the `Stoch'. Our goal is to realize a robust, modular, and a reliable quadrupedal platform, using which various locomotion behaviors are explored. This platform enables us to explore different research problems in legged locomotion, which use both traditional and learning based techniques. We discuss the merits and limitations of the platform in terms of exploitation of available behaviours, fast rapid prototyping, reproduction and repair. Towards the end, we will demonstrate trotting, bounding behaviors, and preliminary results in turning. In addition, we will also show various gait transitions i.e., trot-to-turn and trot-to-bound behaviors. |
| 15. | K P, Naveen; Rajesh Sundaresan Double-Auction Mechanisms for Resource Trading Markets Journal Article arXiv: Computer Science, 2019. Abstract | BibTeX | Tags: Networking and Internet Architecture | Links: @article{Naveen2019, title = {Double-Auction Mechanisms for Resource Trading Markets}, author = {K P, Naveen; Rajesh Sundaresan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/Double-auction-Mechanisms-for-resource-trading-markets.pdf}, year = {2019}, date = {2019-01-27}, journal = {arXiv: Computer Science}, abstract = {We consider a double-auction mechanism, which was recently proposed in the context of a mobile data-offloading market. It is also applicable in a network slicing market. Network operators (users) derive benefit from offloading their traffic to third party WiFi or femtocell networks (link-suppliers). Link-suppliers experience costs for the additional capacity that they provide. Users and link-suppliers (collectively referred to as agents) have their pay-offs and cost functions as private knowledge. A system-designer decomposes the problem into a network problem (with surrogate pay-offs and surrogate cost functions) and agent problems (one per agent). The surrogate pay-offs and cost functions are modulated by the agents’ bids. Agents’ payoffs and costs are then determined by the allocations and prices set by the system designer. Under this design, so long as the agents do not anticipate the effect of their actions, a competitive equilibrium exists as a solution to the network and agent problems, and this equilibrium optimizes the system utility. However, this design fails when the agents are strategic (price anticipating). The presence of strategic supplying agents drives the system to an undesirable equilibrium with zero participation resulting in an efficiency loss of 100%. This is in stark contrast to the setting when link-suppliers are not strategic: the efficiency loss is at most 34% when the users alone are strategic. The paper then proposes a Stackelberg game modification with asymmetric information structures for suppliers and users in order to alleviate the efficiency loss problem. The system designer first announces the allocation and payment functions. He then invites the supplying agents to announce their bids, following which the users are invited to respond to the suppliers’ bids. The resulting Stackelberg games efficiency losses can be characterized in terms of the suppliers cost functions when the user pay-off functions are linear. Specifically, when the link-supplier’s cost function is quadratic, the worst case efficiency loss is 25%. Further, the loss in efficiency improves for polynomial cost functions of higher degree.}, keywords = {Networking and Internet Architecture}, pubstate = {published}, tppubtype = {article} } We consider a double-auction mechanism, which was recently proposed in the context of a mobile data-offloading market. It is also applicable in a network slicing market. Network operators (users) derive benefit from offloading their traffic to third party WiFi or femtocell networks (link-suppliers). Link-suppliers experience costs for the additional capacity that they provide. Users and link-suppliers (collectively referred to as agents) have their pay-offs and cost functions as private knowledge. A system-designer decomposes the problem into a network problem (with surrogate pay-offs and surrogate cost functions) and agent problems (one per agent). The surrogate pay-offs and cost functions are modulated by the agents’ bids. Agents’ payoffs and costs are then determined by the allocations and prices set by the system designer. Under this design, so long as the agents do not anticipate the effect of their actions, a competitive equilibrium exists as a solution to the network and agent problems, and this equilibrium optimizes the system utility. However, this design fails when the agents are strategic (price anticipating). The presence of strategic supplying agents drives the system to an undesirable equilibrium with zero participation resulting in an efficiency loss of 100%. This is in stark contrast to the setting when link-suppliers are not strategic: the efficiency loss is at most 34% when the users alone are strategic. The paper then proposes a Stackelberg game modification with asymmetric information structures for suppliers and users in order to alleviate the efficiency loss problem. The system designer first announces the allocation and payment functions. He then invites the supplying agents to announce their bids, following which the users are invited to respond to the suppliers’ bids. The resulting Stackelberg games efficiency losses can be characterized in terms of the suppliers cost functions when the user pay-off functions are linear. Specifically, when the link-supplier’s cost function is quadratic, the worst case efficiency loss is 25%. Further, the loss in efficiency improves for polynomial cost functions of higher degree. |
| 16. | Akshatha, Jagadish; Manoj Varma Enhancing the capture efficiency of a micro-robotic swarm in targeted delivery using velocity control Conference MARSS – Manipulation, Automation and Robotics at Small Scale, 2019. Abstract | BibTeX | Tags: Robot swarms | Links: @conference{Akshatha2019, title = {Enhancing the capture efficiency of a micro-robotic swarm in targeted delivery using velocity control}, author = {Akshatha, Jagadish; Manoj Varma}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/Enhancing-the-capture-efficiency-of-a-micro-robotic-swarm-in-targeted-delivery-using-velocity-control.pdf}, year = {2019}, date = {2019-07-05}, booktitle = {MARSS – Manipulation, Automation and Robotics at Small Scale}, journal = {MARSS – Manipulation, Automation and Robotics at Small Scale 2019}, abstract = {Swarm robotics at the micro scale is one of the popular research areas currently. However, there has been no generic approach to modelling, designing and controlling these micro-robotic swarms. In this paper, we investigate the capture efficiency of a micro-robotic swarm for the specific application of targeted delivery. We mathematically model the natural dynamics of the swarm that would help us understand the inherent passive behaviour of the swarm system. This would lead to optimal design and control of the micro-bots. We also investigate the effect of the mean control of the swarm on the capture efficiency, which describes the efficiency of the microrobotic swarm in reaching the target.}, keywords = {Robot swarms}, pubstate = {published}, tppubtype = {conference} } Swarm robotics at the micro scale is one of the popular research areas currently. However, there has been no generic approach to modelling, designing and controlling these micro-robotic swarms. In this paper, we investigate the capture efficiency of a micro-robotic swarm for the specific application of targeted delivery. We mathematically model the natural dynamics of the swarm that would help us understand the inherent passive behaviour of the swarm system. This would lead to optimal design and control of the micro-bots. We also investigate the effect of the mean control of the swarm on the capture efficiency, which describes the efficiency of the microrobotic swarm in reaching the target. |
| 17. | Pavankumar Tallapragada; Massimo Franceschetti; Jorge, Cortes Event-triggered control under time-varying rates and channel blackouts Journal Article {IFAC} Journal of Systems and Control , 9 , 2019. Abstract | BibTeX | Tags: Control and Optimisation, Networked Control systems | Links: @article{@Pavankumar2019b, title = {Event-triggered control under time-varying rates and channel blackouts}, author = {Pavankumar Tallapragada; Massimo Franceschetti; Jorge, Cortes}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Event-tiggered-control-under-time-varying-rates-and-channel-blackouts.pdf}, doi = {10.1016/j.ifacsc.2019.100064}, year = {2019}, date = {2019-09-30}, journal = {{IFAC} Journal of Systems and Control }, volume = {9}, abstract = {This paper studies event-triggered stabilization of linear time-invariant systems over time-varying rate-limited communication channels. We explicitly account for the possibility of channel blackouts, i.e., intervals of time when the communication channel is unavailable for feedback. We assume prior knowledge of the channel evolution and we introduce the notion of bit capacity as the maximum total number of bits that could be communicated over a given time interval. We then provide an efficient real-time algorithm to lower bound the bit capacity for a deterministic channel whose characteristics are piecewise constant in time. Building on these results, we design an event-triggering strategy that guarantees Zeno-free, exponential stabilization at a desired convergence rate even in the presence of intermittent channel blackouts. The contributions are the notion of channel blackouts, the effective event-triggered control despite their occurrence, and the analysis and quantification of the bit capacity for a class of time-varying continuous-time channels. Various simulations illustrate the results.}, keywords = {Control and Optimisation, Networked Control systems}, pubstate = {published}, tppubtype = {article} } This paper studies event-triggered stabilization of linear time-invariant systems over time-varying rate-limited communication channels. We explicitly account for the possibility of channel blackouts, i.e., intervals of time when the communication channel is unavailable for feedback. We assume prior knowledge of the channel evolution and we introduce the notion of bit capacity as the maximum total number of bits that could be communicated over a given time interval. We then provide an efficient real-time algorithm to lower bound the bit capacity for a deterministic channel whose characteristics are piecewise constant in time. Building on these results, we design an event-triggering strategy that guarantees Zeno-free, exponential stabilization at a desired convergence rate even in the presence of intermittent channel blackouts. The contributions are the notion of channel blackouts, the effective event-triggered control despite their occurrence, and the analysis and quantification of the bit capacity for a class of time-varying continuous-time channels. Various simulations illustrate the results. |
| 18. | Shourya Bose; Pavankumar, Tallapragada Event-Triggered Second Moment Stabilization under Markov Packet Drops Conference 2019 Fifth Indian Control Conference (ICC), 9-11 Jan. 2019, New Delhi, India, 2019. Abstract | BibTeX | Tags: Control and Optimisation, Networked Control systems | Links: @conference{@Shourya2019, title = {Event-Triggered Second Moment Stabilization under Markov Packet Drops}, author = {Shourya Bose; Pavankumar,Tallapragada}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Event-triggered-second-moment-stabilization-under-Markov-Packet-drops.pdf}, doi = {10.1109/INDIANCC.2019.8715576}, year = {2019}, date = {2019-05-16}, booktitle = {2019 Fifth Indian Control Conference (ICC), 9-11 Jan. 2019, New Delhi, India}, journal = {Fifth Indian Control Conference (ICC), 9-11 Jan. 2019, New Delhi, India}, pages = {113-118}, abstract = {In this paper, we consider the problem of second moment stabilization of scalar linear systems under Markov packet drops. We assume that the channel state evolution is given by a Markov chain. Corresponding to each possible value of the channel state the packet drop probability distribution may be different. We design an event-triggered transmission policy that ensures that the second moment of the plant state converges exponentially, at a desired rate, to an ultimate bound. Our approach relies on an online assessment of the necessity of transmission at any given time step and a transmission occurs only if necessary. We illustrate the results through simulations.}, keywords = {Control and Optimisation, Networked Control systems}, pubstate = {published}, tppubtype = {conference} } In this paper, we consider the problem of second moment stabilization of scalar linear systems under Markov packet drops. We assume that the channel state evolution is given by a Markov chain. Corresponding to each possible value of the channel state the packet drop probability distribution may be different. We design an event-triggered transmission policy that ensures that the second moment of the plant state converges exponentially, at a desired rate, to an ultimate bound. Our approach relies on an online assessment of the necessity of transmission at any given time step and a transmission occurs only if necessary. We illustrate the results through simulations. |
| 19. | Eduardo Hargreavesa; Claudio Agosti; Daniel Menasché; Giovanni Neglia; Alexandre Reiffers-Masson; Eitan, Altman Fairness in Online Social Network Timelines: Measurements, Models and Mechanism Design Journal Article IFIP Perfomance 2018 - 36th International Symposium on Computer Performance, Modeling, Measurements and Evaluation, 129 , pp. 15-39, 2019. Abstract | BibTeX | Tags: measuremets | Links: @article{@Alex2019, title = {Fairness in Online Social Network Timelines: Measurements, Models and Mechanism Design}, author = {Eduardo Hargreavesa; Claudio Agosti; Daniel Menasché; Giovanni Neglia; Alexandre Reiffers-Masson; Eitan, Altman}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/11/Fairness-in-Online-Social-Network-Timelines-Measurements-Models-and-Mechanism-Design-Journal-Article.pdf}, doi = {10.1016/j.peva.2018.09.009}, year = {2019}, date = {2019-02-01}, journal = {IFIP Perfomance 2018 - 36th International Symposium on Computer Performance, Modeling, Measurements and Evaluation}, volume = {129}, pages = {15-39}, abstract = {Facebook News Feed personalization algorithm has a significant impact, on a daily basis, on the lifestyle, mood and opinion of millions of Internet users. Nonetheless, the behavior of such algorithm lacks transparency, motivating measurements, modeling and analysis in order to understand and improve its properties. In this paper, we propose a reproducible methodology encompassing measurements, an analytical model and a fairness-based News Feed design. The model leverages the versatility and analytical tractability of time-to-live (TTL) counters to capture the visibility and occupancy of publishers over a News Feed. Measurements are used to parameterize and to validate the expressive power of the proposed model. Then, we conduct a what–if analysis to assess the visibility and occupancy bias incurred by users against a baseline derived from the model. Our results indicate that a significant bias exists and it is more prominent at the top position of the News Feed. In addition, we find that the bias is non-negligible even for users that are deliberately set as neutral with respect to their political views, motivating the proposal of a novel and more transparent fairness-based News Feed design.}, keywords = {measuremets}, pubstate = {published}, tppubtype = {article} } Facebook News Feed personalization algorithm has a significant impact, on a daily basis, on the lifestyle, mood and opinion of millions of Internet users. Nonetheless, the behavior of such algorithm lacks transparency, motivating measurements, modeling and analysis in order to understand and improve its properties. In this paper, we propose a reproducible methodology encompassing measurements, an analytical model and a fairness-based News Feed design. The model leverages the versatility and analytical tractability of time-to-live (TTL) counters to capture the visibility and occupancy of publishers over a News Feed. Measurements are used to parameterize and to validate the expressive power of the proposed model. Then, we conduct a what–if analysis to assess the visibility and occupancy bias incurred by users against a baseline derived from the model. Our results indicate that a significant bias exists and it is more prominent at the top position of the News Feed. In addition, we find that the bias is non-negligible even for users that are deliberately set as neutral with respect to their political views, motivating the proposal of a novel and more transparent fairness-based News Feed design. |
| 20. | Babu, Arun; Thomas, Jithin Jose Freestyle, a randomized version of ChaCha for resisting offline brute-force and dictionary attacks Journal Article Forthcoming Journal of Information Security and Applications, 49 , Forthcoming. Abstract | BibTeX | Tags: Smart City | Links: @article{Babu2019, title = {Freestyle, a randomized version of ChaCha for resisting offline brute-force and dictionary attacks}, author = {Arun Babu and Jithin Jose Thomas}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Freestyle-a-randomized-version-of-ChaCha-for-resisting-offline-brute-force-and-dictionary-attacks.pdf}, doi = {10.1016/j.jisa.2019.102396}, year = {2019}, date = {2019-12-01}, journal = {Journal of Information Security and Applications}, volume = {49}, abstract = {This paper introduces Freestyle, a randomized and variable round version of the ChaCha cipher. Freestyle uses the concept of hash based halting condition where a decryption attempt with an incorrect key is likely to take longer time to halt. This makes Freestyle resistant to key-guessing attacks i.e. brute-force and dictionary based attacks. Freestyle demonstrates a novel approach for ciphertext randomization by using random number of rounds for each block, where the exact number of rounds are unknown to the receiver in advance. Freestyle provides the possibility of generating 2^128 different ciphertexts for a given key, nonce, and message; thus resisting key and nonce reuse attacks. Due to its inherent random behavior, Freestyle makes cryptanalysis through known-plaintext, chosen-plaintext, and chosen-ciphertext attacks difficult in practice. On the other hand, Freestyle has costlier cipher initialization process, typically generates 3.125% larger ciphertext, and was found to be 1.6 to 3.2 times slower than ChaCha20. Freestyle is suitable for applications that favor ciphertext randomization and resistance to key-guessing and key reuse attacks over performance and ciphertext size. Freestyle is ideal for applications where ciphertext can be assumed to be in full control of an adversary, and an offline key-guessing attack can be carried out. }, keywords = {Smart City}, pubstate = {forthcoming}, tppubtype = {article} } This paper introduces Freestyle, a randomized and variable round version of the ChaCha cipher. Freestyle uses the concept of hash based halting condition where a decryption attempt with an incorrect key is likely to take longer time to halt. This makes Freestyle resistant to key-guessing attacks i.e. brute-force and dictionary based attacks. Freestyle demonstrates a novel approach for ciphertext randomization by using random number of rounds for each block, where the exact number of rounds are unknown to the receiver in advance. Freestyle provides the possibility of generating 2^128 different ciphertexts for a given key, nonce, and message; thus resisting key and nonce reuse attacks. Due to its inherent random behavior, Freestyle makes cryptanalysis through known-plaintext, chosen-plaintext, and chosen-ciphertext attacks difficult in practice. On the other hand, Freestyle has costlier cipher initialization process, typically generates 3.125% larger ciphertext, and was found to be 1.6 to 3.2 times slower than ChaCha20. Freestyle is suitable for applications that favor ciphertext randomization and resistance to key-guessing and key reuse attacks over performance and ciphertext size. Freestyle is ideal for applications where ciphertext can be assumed to be in full control of an adversary, and an offline key-guessing attack can be carried out. |
| 21. | Pavankumar Tallapragada; Jorge, Cortes Hierarchical-distributed optimized coordination of intersection traffic Journal Article IEEE Transactions on Intelligent Transportation Systems, pp. 1-14, 2019. Abstract | BibTeX | Tags: Control and Optimisation, Networked Tansportation | Links: @article{@Pavankumar2019, title = {Hierarchical-distributed optimized coordination of intersection traffic}, author = {Pavankumar Tallapragada; Jorge,Cortes}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Hierarchical-distributed-optimized-coordination-of-intersection-traffic.pdf}, doi = {10.1109/TITS.2019.2912881}, year = {2019}, date = {2019-05-08}, journal = {IEEE Transactions on Intelligent Transportation Systems}, pages = {1-14}, abstract = {This paper considers the problem of coordinating vehicular traffic at an intersection and on the branches leading to it in order to minimize a combination of total travel time and energy consumption. We propose a provably safe hierarchical-distributed solution to balance computational complexity and optimality of the system operation. In our design, a central intersection manager communicates with vehicles heading toward the intersection, groups them into clusters (termed bubbles) as they appear, and determines an optimal schedule of passage through the intersection for each bubble. The vehicles in each bubble receive their schedule and implement local distributed control to ensure system-wide inter-vehicular safety while respecting speed and acceleration limits, conforming to the assigned schedule, and seeking to optimize their individual trajectories. Our analysis rigorously establishes that the different aspects of the hierarchical design operate in concert and that the safety specifications are satisfied. We illustrate its execution through a suite of simulations and compare its performance against optimized signal-based coordination over a wide range of system parameters.}, keywords = {Control and Optimisation, Networked Tansportation}, pubstate = {published}, tppubtype = {article} } This paper considers the problem of coordinating vehicular traffic at an intersection and on the branches leading to it in order to minimize a combination of total travel time and energy consumption. We propose a provably safe hierarchical-distributed solution to balance computational complexity and optimality of the system operation. In our design, a central intersection manager communicates with vehicles heading toward the intersection, groups them into clusters (termed bubbles) as they appear, and determines an optimal schedule of passage through the intersection for each bubble. The vehicles in each bubble receive their schedule and implement local distributed control to ensure system-wide inter-vehicular safety while respecting speed and acceleration limits, conforming to the assigned schedule, and seeking to optimize their individual trajectories. Our analysis rigorously establishes that the different aspects of the hierarchical design operate in concert and that the safety specifications are satisfied. We illustrate its execution through a suite of simulations and compare its performance against optimized signal-based coordination over a wide range of system parameters. |
| 22. | Aadirupa, Saha; Rakesh Shivanna; Chiranjib Bhattacharyya How Many Pairwise Preferences Do We Need to Rank a Graph Consistently? Conference Proceedings of the AAAI Conference on Artificial Intelligence, 33 , 2019. Abstract | BibTeX | Tags: Machine Learning | Links: @conference{Aadirupa2019, title = {How Many Pairwise Preferences Do We Need to Rank a Graph Consistently?}, author = {Aadirupa, Saha; Rakesh Shivanna; Chiranjib Bhattacharyya}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/How-many-pairwise-preferences-do-we-need-to-rank-a-graph-consistently-1.pdf}, doi = {10.1609/aaai.v33i01.33014830}, year = {2019}, date = {2019-07-17}, booktitle = {Proceedings of the AAAI Conference on Artificial Intelligence}, volume = {33}, pages = {4830-4837}, abstract = {We consider the problem of optimal recovery of true ranking of n items from a randomly chosen subset of their pairwise preferences. It is well known that without any further assumption, one requires a sample size of Ω(n2) for the purpose. We analyze the problem with an additional structure of relational graph G([n],E) over the n items added with an assumption of locality: Neighboring items are similar in their rankings. Noting the preferential nature of the data, we choose to embed not the graph, but, its strong product to capture the pairwise node relationships. Furthermore, unlike existing literature that uses Laplacian embedding for graph based learning problems, we use a richer class of graph embeddings—orthonormal representations—that includes (normalized) Laplacian as its special case. Our proposed algorithm, Pref-Rank, predicts the underlying ranking using an SVM based approach using the chosen embedding of the product graph, and is the first to provide statistical consistency on two ranking losses: Kendall’s tau and Spearman’s footrule, with a required sample complexity of O(n2χ(G¯))⅔ pairs, χ(G¯) being the chromatic number of the complement graph G¯. Clearly, our sample complexity is smaller for dense graphs, with χ(G¯) characterizing the degree of node connectivity, which is also intuitive due to the locality assumption e.g. O(n4/3) for union of k-cliques, or O(n5/3) for random and power law graphs etc.—a quantity much smaller than the fundamental limit of Ω(n2) for large n. This, for the first time, relates ranking complexity to structural properties of the graph. We also report experimental evaluations on different synthetic and real-world datasets, where our algorithm is shown to outperform the state of the art methods.}, keywords = {Machine Learning}, pubstate = {published}, tppubtype = {conference} } We consider the problem of optimal recovery of true ranking of n items from a randomly chosen subset of their pairwise preferences. It is well known that without any further assumption, one requires a sample size of Ω(n2) for the purpose. We analyze the problem with an additional structure of relational graph G([n],E) over the n items added with an assumption of locality: Neighboring items are similar in their rankings. Noting the preferential nature of the data, we choose to embed not the graph, but, its strong product to capture the pairwise node relationships. Furthermore, unlike existing literature that uses Laplacian embedding for graph based learning problems, we use a richer class of graph embeddings—orthonormal representations—that includes (normalized) Laplacian as its special case. Our proposed algorithm, Pref-Rank, predicts the underlying ranking using an SVM based approach using the chosen embedding of the product graph, and is the first to provide statistical consistency on two ranking losses: Kendall’s tau and Spearman’s footrule, with a required sample complexity of O(n2χ(G¯))⅔ pairs, χ(G¯) being the chromatic number of the complement graph G¯. Clearly, our sample complexity is smaller for dense graphs, with χ(G¯) characterizing the degree of node connectivity, which is also intuitive due to the locality assumption e.g. O(n4/3) for union of k-cliques, or O(n5/3) for random and power law graphs etc.—a quantity much smaller than the fundamental limit of Ω(n2) for large n. This, for the first time, relates ranking complexity to structural properties of the graph. We also report experimental evaluations on different synthetic and real-world datasets, where our algorithm is shown to outperform the state of the art methods. |
| 23. | Suran, Swathi; Varma, Manoj Imaging flow distribution through nanoporous polymer films using bright-field nanoscopy Journal Article Journal of Applied Physics, 125 (16), pp. 165303:1-5, 2019. Abstract | BibTeX | Tags: Sensing and Actuation Systems | Links: @article{Suran2019, title = {Imaging flow distribution through nanoporous polymer films using bright-field nanoscopy }, author = {Swathi Suran and Manoj Varma}, url = {http://www.rbccps.org/wp-content/uploads/2019/05/1.5089278.pdf}, doi = {10.1063/1.5089278}, year = {2019}, date = {2019-04-23}, journal = {Journal of Applied Physics}, volume = {125}, number = {16}, pages = {165303:1-5}, abstract = {Bright-field nanoscopy (BFN), an optical visualization technique using standard diffraction-limited microscopy, has been used in the past to visualize nanoscale objects such as grain boundaries in single layer graphene and water transport across nanomembranes. Here, we show how BFN can be used to visualize and gather information about water transport networks in porous polymer membranes. We were able to visualize a sub-100 nm network of pores in a polyelectrolyte multilayer using this technique. The extracted pore diameter distribution fitted well to an exponential distribution. Further, the time evolution of pore diameter displayed two distinct regimes consistent with expectations.}, keywords = {Sensing and Actuation Systems}, pubstate = {published}, tppubtype = {article} } Bright-field nanoscopy (BFN), an optical visualization technique using standard diffraction-limited microscopy, has been used in the past to visualize nanoscale objects such as grain boundaries in single layer graphene and water transport across nanomembranes. Here, we show how BFN can be used to visualize and gather information about water transport networks in porous polymer membranes. We were able to visualize a sub-100 nm network of pores in a polyelectrolyte multilayer using this technique. The extracted pore diameter distribution fitted well to an exponential distribution. Further, the time evolution of pore diameter displayed two distinct regimes consistent with expectations. |
| 24. | Shounak Bhattacharya Abhik Singla, Abhimanyu Singh Dhaivat Dholakiya Shalabh Bhatnagar Bharadwaj Amrutur Ashitava Ghosal Shishir Kolathaya Learning Active Spine Behaviors for Dynamic and Efficient Locomotion in Quadruped Robots Conference The 28th IEEE International Conference on Robot & Human Interactive Communication, 2019. Abstract | BibTeX | Tags: Quadruped, Robotics, Spine | Links: @conference{Bhattacharya2019, title = {Learning Active Spine Behaviors for Dynamic and Efficient Locomotion in Quadruped Robots}, author = {Shounak Bhattacharya, Abhik Singla, Abhimanyu Singh, Dhaivat Dholakiya, Shalabh Bhatnagar, Bharadwaj Amrutur, Ashitava Ghosal, Shishir Kolathaya}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/Learning-Active-Spine-Behaviors-for-Dynamic-and-efficient-Locomotion-in-quadruped-Robots.pdf}, year = {2019}, date = {2019-10-14}, booktitle = {The 28th IEEE International Conference on Robot & Human Interactive Communication}, abstract = {Abstract— In this work, we provide a simulation framework to perform systematic studies on the effects of spinal joint compliance and actuation on bounding performance of a 16-DOF quadruped spined robot Stoch 2. Fast quadrupedal locomotion with active spine is an extremely hard problem, and involves a complex coordination between the various degrees of freedom. Therefore, past attempts at addressing this problem have not seen much success. Deep-Reinforcement Learning seems to be a promising approach, after its recent success in a variety of robot platforms, and the goal of this paper is to use this approach to realize the aforementioned behaviors. With this learning framework, the robot reached a bounding speed of 2.1 m/s with a maximum Froude number of 2. Simulation results also show that use of active spine, indeed, increased the stride length, improved the cost of transport, and also reduced the natural frequency to more realistic values. We will support these results with the simulation videos.}, keywords = {Quadruped, Robotics, Spine}, pubstate = {published}, tppubtype = {conference} } Abstract— In this work, we provide a simulation framework to perform systematic studies on the effects of spinal joint compliance and actuation on bounding performance of a 16-DOF quadruped spined robot Stoch 2. Fast quadrupedal locomotion with active spine is an extremely hard problem, and involves a complex coordination between the various degrees of freedom. Therefore, past attempts at addressing this problem have not seen much success. Deep-Reinforcement Learning seems to be a promising approach, after its recent success in a variety of robot platforms, and the goal of this paper is to use this approach to realize the aforementioned behaviors. With this learning framework, the robot reached a bounding speed of 2.1 m/s with a maximum Froude number of 2. Simulation results also show that use of active spine, indeed, increased the stride length, improved the cost of transport, and also reduced the natural frequency to more realistic values. We will support these results with the simulation videos. |
| 25. | Gayathri R, Prabhu; Srikrishna Bhashyam; Aditya Gopalan; Rajesh Sundaresan Learning to detect an anomalous target with observations from an exponential family Conference 2019 IEEE Data Science Workshop (DSW), Minneapolis, USA, 2-5 June , 2019. Abstract | BibTeX | Tags: Electrical Communications Engineering | Links: @conference{@Gayathri2019, title = {Learning to detect an anomalous target with observations from an exponential family}, author = {Gayathri, R, Prabhu; Srikrishna Bhashyam; Aditya Gopalan; Rajesh, Sundaresan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Learning-to-detect-an-anomalous-target-with-observation-from-an-exponential-family.pdf}, doi = {10.1109/DSW.2019.8755796}, year = {2019}, date = {2019-07-04}, booktitle = {2019 IEEE Data Science Workshop (DSW), Minneapolis, USA, 2-5 June }, abstract = {The problem of identifying an anomalous arm from a set of K arms, with fixed confidence, is studied in a sequential decision-making scenario. Each arm's signal follows a distribution from the vector parameter exponential family. The actual parameters of the anomalous and regular arms are unknown. Further, the decision maker incurs a cost for switching from one arm to another. A sequential policy based on a modified generalised likelihood ratio statistic is proposed. The policy, with a suitable threshold, is shown to satisfy the given constraint on the probability of false detection. Further, the proposed policy is asymptotically optimal in terms of the total cost among all policies that satisfy the constraint on the probability of false detection.}, keywords = {Electrical Communications Engineering}, pubstate = {published}, tppubtype = {conference} } The problem of identifying an anomalous arm from a set of K arms, with fixed confidence, is studied in a sequential decision-making scenario. Each arm's signal follows a distribution from the vector parameter exponential family. The actual parameters of the anomalous and regular arms are unknown. Further, the decision maker incurs a cost for switching from one arm to another. A sequential policy based on a modified generalised likelihood ratio statistic is proposed. The policy, with a suitable threshold, is shown to satisfy the given constraint on the probability of false detection. Further, the proposed policy is asymptotically optimal in terms of the total cost among all policies that satisfy the constraint on the probability of false detection. |
| 26. | P N, Karthik; Rajesh Sundaresan Learning to Detect an Odd Markov Arm Journal Article arXiv: Computer Science, 2019. Abstract | BibTeX | Tags: Information Theory | Links: @article{Karthik2019, title = {Learning to Detect an Odd Markov Arm}, author = {P N, Karthik; Rajesh Sundaresan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/Learning-to-detect-an-odd-markov-arm.pdf}, year = {2019}, date = {2019-05-01}, journal = {arXiv: Computer Science}, abstract = {A multi-armed bandit with finitely many arms is studied when each arm is a homogeneous Markov process on an underlying finite state space. The transition law of one of the arms, referred to as the odd arm, is different from the common transition law of all other arms. A learner, who has no knowledge of the above transition laws, has to devise a sequential test to identify the index of the odd arm as quickly as possible, subject to an upper bound on the probability of error. For this problem, we derive an asymptotic lower bound on the expected stopping time of any sequential test of the learner, where the asymptotics is as the probability of error vanishes. Furthermore, we propose a sequential test, and show that the asymptotic behaviour of its expected stopping time comes arbitrarily close to that of the lower bound. Prior works deal with iid arms, whereas our work deals with Markov arms. Our analysis of the rested Markov setting is a key first step in understanding the difficult case of restless Markov setting, which is still open.}, keywords = {Information Theory}, pubstate = {published}, tppubtype = {article} } A multi-armed bandit with finitely many arms is studied when each arm is a homogeneous Markov process on an underlying finite state space. The transition law of one of the arms, referred to as the odd arm, is different from the common transition law of all other arms. A learner, who has no knowledge of the above transition laws, has to devise a sequential test to identify the index of the odd arm as quickly as possible, subject to an upper bound on the probability of error. For this problem, we derive an asymptotic lower bound on the expected stopping time of any sequential test of the learner, where the asymptotics is as the probability of error vanishes. Furthermore, we propose a sequential test, and show that the asymptotic behaviour of its expected stopping time comes arbitrarily close to that of the lower bound. Prior works deal with iid arms, whereas our work deals with Markov arms. Our analysis of the rested Markov setting is a key first step in understanding the difficult case of restless Markov setting, which is still open. |
| 27. | Soundararajan, Rajiv; Biswas, Soma Machine vision quality assessment for robust face detection Journal Article Signal Processing: Image Communication, 72 , pp. 92-104, 2019. Abstract | BibTeX | Tags: Student Research Grant | Links: @article{Soundararajan2019, title = {Machine vision quality assessment for robust face detection}, author = {Rajiv Soundararajan and Soma Biswas}, url = {http://www.rbccps.org/wp-content/uploads/2019/05/1-s2.0-S0923596518306957-main.pdf}, doi = {10.1016/j.image.2018.12.012}, year = {2019}, date = {2019-03-29}, journal = {Signal Processing: Image Communication}, volume = {72}, pages = {92-104}, abstract = {Machine vision quality (MVQ) assessment for face detection refers to image quality as judged by face detection algorithms. While perceptual image quality usually depends on human observers, MVQ depends on factors such as the face detection algorithm and performance measures such as recall and precision. We define the MVQ index as a weighted combination of the predicted precision and recall of the face detection algorithm and predict it using image features based on natural scene statistics. A filter bank framework is developed to achieve image enhancement of distorted images for face detection, where the enhancement operations can be optimized for either perceptual quality or MVQ. It is shown that MVQ can be effectively used to control the relative performance of recall and precision that can be achieved on enhanced images. For certain face detection algorithms, optimizing for MVQ instead of perceptual quality can lead to improved face detection performance. The MVQ is developed for three different face detection algorithms and the image enhancement framework is tested on a dataset based on the IDEAL-LIVE Distorted Face Database. A computationally efficient method to optimize for MVQ is designed by predicting the MVQ of enhanced images directly from features extracted from distorted images and filter parameters. The optimization framework is further enhanced by allowing for filter bank independent optimization of MVQ.}, keywords = {Student Research Grant}, pubstate = {published}, tppubtype = {article} } Machine vision quality (MVQ) assessment for face detection refers to image quality as judged by face detection algorithms. While perceptual image quality usually depends on human observers, MVQ depends on factors such as the face detection algorithm and performance measures such as recall and precision. We define the MVQ index as a weighted combination of the predicted precision and recall of the face detection algorithm and predict it using image features based on natural scene statistics. A filter bank framework is developed to achieve image enhancement of distorted images for face detection, where the enhancement operations can be optimized for either perceptual quality or MVQ. It is shown that MVQ can be effectively used to control the relative performance of recall and precision that can be achieved on enhanced images. For certain face detection algorithms, optimizing for MVQ instead of perceptual quality can lead to improved face detection performance. The MVQ is developed for three different face detection algorithms and the image enhancement framework is tested on a dataset based on the IDEAL-LIVE Distorted Face Database. A computationally efficient method to optimize for MVQ is designed by predicting the MVQ of enhanced images directly from features extracted from distorted images and filter parameters. The optimization framework is further enhanced by allowing for filter bank independent optimization of MVQ. |
| 28. | D Thirumulanathan; Rajesh Sundaresan; Y, Narahari On Optimal Mechanisms in the Two-Item Single-Buyer Unit-Demand Setting Journal Article arXiv: Computer Science, 2019. Abstract | BibTeX | Tags: Electrical Communications Engineering | Links: @article{Thirumulanathan2019, title = {On Optimal Mechanisms in the Two-Item Single-Buyer Unit-Demand Setting}, author = {D, Thirumulanathan; Rajesh Sundaresan; Y, Narahari}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/On-Optimal-Mechanisms-in-the-Two-Item-Single-Buyer-Unit-Demand-Setting.pdf}, year = {2019}, date = {2019-03-05}, journal = {arXiv: Computer Science}, abstract = {We consider the problem of designing a revenue-optimal mechanism in the two-item, single-buyer, unit-demand setting when the buyer’s valuations, (z1, z2), are uniformly distributed in an arbitrary rectangle [c, c+b1]×[c, c+b2] in the positive quadrant. We provide a complete and explicit solution for arbitrary nonnegative values of (c, b1, b2). We identify five simple structures, each with at most five (possibly stochastic) menu items, and prove that the optimal mechanism has one of the five structures. We also characterize the optimal mechanism as a function of b1, b2, and c. When c is low, the optimal mechanism is a posted price mechanism with an exclusion region; when c is high, it is a posted price mechanism without an exclusion region. Our results are the first to show the existence of optimal mechanisms with no exclusion region, to the best of our knowledge.}, keywords = {Electrical Communications Engineering}, pubstate = {published}, tppubtype = {article} } We consider the problem of designing a revenue-optimal mechanism in the two-item, single-buyer, unit-demand setting when the buyer’s valuations, (z1, z2), are uniformly distributed in an arbitrary rectangle [c, c+b1]×[c, c+b2] in the positive quadrant. We provide a complete and explicit solution for arbitrary nonnegative values of (c, b1, b2). We identify five simple structures, each with at most five (possibly stochastic) menu items, and prove that the optimal mechanism has one of the five structures. We also characterize the optimal mechanism as a function of b1, b2, and c. When c is low, the optimal mechanism is a posted price mechanism with an exclusion region; when c is high, it is a posted price mechanism without an exclusion region. Our results are the first to show the existence of optimal mechanisms with no exclusion region, to the best of our knowledge. |
| 29. | Subhajit Goswami; Pavankumar, Tallapragada One-Shot Coordination of Feeder Vehicles for Multi-Modal Transportation Conference 18th European Control Conference (ECC), Naples, Italy, 25-28 June, 2019. Abstract | BibTeX | Tags: Control and Optimisation, Networked Tansportation | Links: @conference{@Subhajit2019, title = {One-Shot Coordination of Feeder Vehicles for Multi-Modal Transportation}, author = {Subhajit Goswami; Pavankumar,Tallapragada}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/One-shot-coordination-of-feeder-vehicles-for-multi-modal-transportation.pdf}, doi = {10.23919/ECC.2019.8795701}, year = {2019}, date = {2019-08-15}, booktitle = {18th European Control Conference (ECC), Naples, Italy, 25-28 June}, pages = {1714-1719}, abstract = {In this paper, we consider coordinated control of feeder vehicles in the first leg of a multi-modal transport system. In particular, we consider a one-shot problem wherein the passengers must be transported to a common destination before a fixed deadline. We pose the problem as the maximization of the service provider's profits through optimal pricing and feeder allocations given the knowledge of demand and supply distributions. Though the original problem is non-linear, with optimal pricing scheme we reduce it to a linear program. We then move on to developing an off-line route elimination algorithm that reduces the problem size in terms of computation memory and time. Further, we provide a simplified way to compute the maximum possible profits as a function of total supply for a given demand distribution. We test the framework with simulations on a 20 node graph under a fixed-demand profile.}, keywords = {Control and Optimisation, Networked Tansportation}, pubstate = {published}, tppubtype = {conference} } In this paper, we consider coordinated control of feeder vehicles in the first leg of a multi-modal transport system. In particular, we consider a one-shot problem wherein the passengers must be transported to a common destination before a fixed deadline. We pose the problem as the maximization of the service provider's profits through optimal pricing and feeder allocations given the knowledge of demand and supply distributions. Though the original problem is non-linear, with optimal pricing scheme we reduce it to a linear program. We then move on to developing an off-line route elimination algorithm that reduces the problem size in terms of computation memory and time. Further, we provide a simplified way to compute the maximum possible profits as a function of total supply for a given demand distribution. We test the framework with simulations on a 20 node graph under a fixed-demand profile. |
| 30. | Gubbi, Sagar Venkatesh; Amrutur, Bharadwaj One-Shot Object Localization Using Learnt Visual Cues via Siamese Networks Conference IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2019, Macau., 2019. Abstract | BibTeX | Tags: Robotics | Links: @conference{Gubbi2019, title = {One-Shot Object Localization Using Learnt Visual Cues via Siamese Networks}, author = {Sagar Venkatesh Gubbi and Bharadwaj Amrutur}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/09/One-Shot-Object-Localization-using-Learnt-Visual-Cues-via-Siamese-networks.pdf}, year = {2019}, date = {2019-11-04}, booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2019, Macau.}, abstract = {A robot that can operate in novel and unstruc- tured environments must be capable of recognizing new, pre- viously unseen, objects. In this work, a visual cue is used to specify a novel object of interest which must be localized in new environments. An end-to-end neural network equipped with a Siamese network is used to learn the cue, infer the object of interest, and then to localize it in new environments. We show that a simulated robot can pick-and-place novel objects pointed to by a laser pointer. We also evaluate the performance of the proposed approach on a dataset derived from the Omniglot handwritten character dataset and on a small dataset of toys}, keywords = {Robotics}, pubstate = {published}, tppubtype = {conference} } A robot that can operate in novel and unstruc- tured environments must be capable of recognizing new, pre- viously unseen, objects. In this work, a visual cue is used to specify a novel object of interest which must be localized in new environments. An end-to-end neural network equipped with a Siamese network is used to learn the cue, infer the object of interest, and then to localize it in new environments. We show that a simulated robot can pick-and-place novel objects pointed to by a laser pointer. We also evaluate the performance of the proposed approach on a dataset derived from the Omniglot handwritten character dataset and on a small dataset of toys |
| 31. | Vivek Borkar; Alexandre Reiffers, Masson Opinion shaping in social networks using reinforcement learning Journal Article arXiv:1910.08802v1 [cs.SI] , 2019. Abstract | BibTeX | Tags: Social and Information Networks | Links: @article{@Alex2019b, title = {Opinion shaping in social networks using reinforcement learning}, author = {Vivek Borkar; Alexandre Reiffers, Masson}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/11/Opinion-shaping-in-social-networks-using-reinforcement-learning.pdf}, year = {2019}, date = {2019-10-19}, journal = {arXiv:1910.08802v1 [cs.SI] }, abstract = {Recent studies have demonstrated that the decisions of agents in society are shaped by their own intrinsic motivation, and also by the compliance with the social norm. In other words, the decision of acting in a particular manner will be affected by the opinion of society. This social comparison mechanism can lead to imitation behavior, where an agent will try to mimic the behavior of her neighbors. Using this observation, new policies have been designed, e.g., in the context of energy efficiency and transportation choice, to leverage social networks in order to improve altruism and prosocial behavior. One policy is to use targeting strategies. Indeed, by changing the behavior of influential actors in a social network, it is possible to reshape the global behavior of agents towards more prosocial behavior. However, discovering who are the influential agents requires a lot of information, such as the matrix of interactions between agents. In this paper, we study how to shape opinions in social networks when the matrix of interactions is unknown. We consider classical opinion dynamics with some stubborn agents and the possibility of continuously influencing the opinions of a few selected agents, albeit under resource constraints. We map the opinion dynamics to a value iteration scheme for policy evaluation for a specific stochastic shortest path problem. This leads to a representation of the opinion vector as an approximate value function for a stochastic shortest path problem with some non-classical constraints. We suggest two possible ways of influencing agents. One leads to a convex optimization problem and the other to a non-convex one. Firstly, for both problems, we propose two different online two-time scale reinforcement learning schemes that converge to the optimal solution of each problem. Secondly, we suggest stochastic gradient descent schemes and compare these classes of algorithms with the twotime scale reinforcement learning schemes. Thirdly, we also derive another algorithm designed to tackle the curse of dimensionality one faces when all agents are observed. Numerical studies are provided to illustrate the convergence and efficiency of our algorithms.}, keywords = {Social and Information Networks}, pubstate = {published}, tppubtype = {article} } Recent studies have demonstrated that the decisions of agents in society are shaped by their own intrinsic motivation, and also by the compliance with the social norm. In other words, the decision of acting in a particular manner will be affected by the opinion of society. This social comparison mechanism can lead to imitation behavior, where an agent will try to mimic the behavior of her neighbors. Using this observation, new policies have been designed, e.g., in the context of energy efficiency and transportation choice, to leverage social networks in order to improve altruism and prosocial behavior. One policy is to use targeting strategies. Indeed, by changing the behavior of influential actors in a social network, it is possible to reshape the global behavior of agents towards more prosocial behavior. However, discovering who are the influential agents requires a lot of information, such as the matrix of interactions between agents. In this paper, we study how to shape opinions in social networks when the matrix of interactions is unknown. We consider classical opinion dynamics with some stubborn agents and the possibility of continuously influencing the opinions of a few selected agents, albeit under resource constraints. We map the opinion dynamics to a value iteration scheme for policy evaluation for a specific stochastic shortest path problem. This leads to a representation of the opinion vector as an approximate value function for a stochastic shortest path problem with some non-classical constraints. We suggest two possible ways of influencing agents. One leads to a convex optimization problem and the other to a non-convex one. Firstly, for both problems, we propose two different online two-time scale reinforcement learning schemes that converge to the optimal solution of each problem. Secondly, we suggest stochastic gradient descent schemes and compare these classes of algorithms with the twotime scale reinforcement learning schemes. Thirdly, we also derive another algorithm designed to tackle the curse of dimensionality one faces when all agents are observed. Numerical studies are provided to illustrate the convergence and efficiency of our algorithms. |
| 32. | D Thirumulanathana; Rajesh Sundaresan; Y, Narahari Optimal Mechanisms for Selling Two Items to a Single Buyer Having Uniformly Distributed Valuations Journal Article Journal of Mathematical Economics, 82 , pp. 1-30, 2019. Abstract | BibTeX | Tags: Electrical Communications Engineering | Links: @article{@Thirumulanathana, title = {Optimal Mechanisms for Selling Two Items to a Single Buyer Having Uniformly Distributed Valuations}, author = {D,Thirumulanathana; Rajesh Sundaresan; Y,Narahari}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Optimal-Mechanisms-for-Selling-Two-Items-to-a-Single-Buyer-Having-uniformly-distributed-valuations.pdf}, doi = {10.1016/j.jmateco.2019.01.004}, year = {2019}, date = {2019-03-05}, journal = {Journal of Mathematical Economics}, volume = {82}, pages = {1-30}, abstract = {We consider the design of a revenue-optimal mechanism when two items are available to be sold to a single buyer whose valuation is uniformly distributed over an arbitrary rectangle [c1,c1+b1]×[c2,c2+b2] in the positive quadrant. We provide an explicit, complete solution for arbitrary nonnegative values of (c1,c2,b1,b2). We identify eight simple structures, each with at most 4 (possibly stochastic) menu items, and prove that the optimal mechanism has one of these eight structures. We also characterize the optimal mechanism as a function of (c1,c2,b1,b2). The structures indicate that the optimal mechanism involves (a) an interplay of individual sale and a bundle sale when c1 and c2 are low, (b) a bundle sale when c1 and c2 are high, and (c) an individual sale when one of them is high and the other is low. To the best of our knowledge, our results are the first to show the existence of optimal mechanisms with no exclusion region. We further conjecture, based on promising preliminary results, that our methodology can be extended to a wider class of distributions.}, keywords = {Electrical Communications Engineering}, pubstate = {published}, tppubtype = {article} } We consider the design of a revenue-optimal mechanism when two items are available to be sold to a single buyer whose valuation is uniformly distributed over an arbitrary rectangle [c1,c1+b1]×[c2,c2+b2] in the positive quadrant. We provide an explicit, complete solution for arbitrary nonnegative values of (c1,c2,b1,b2). We identify eight simple structures, each with at most 4 (possibly stochastic) menu items, and prove that the optimal mechanism has one of these eight structures. We also characterize the optimal mechanism as a function of (c1,c2,b1,b2). The structures indicate that the optimal mechanism involves (a) an interplay of individual sale and a bundle sale when c1 and c2 are low, (b) a bundle sale when c1 and c2 are high, and (c) an individual sale when one of them is high and the other is low. To the best of our knowledge, our results are the first to show the existence of optimal mechanisms with no exclusion region. We further conjecture, based on promising preliminary results, that our methodology can be extended to a wider class of distributions. |
| 33. | Prafull Prakash; Chaitanya Murti; Saketha Nath; Chiranjib, Bhattacharyya Optimizing DNN architectures for high speed autonomous navigation in GPS denied environments on edge devices Conference Pacific Rim International Conference on Artificial Intelligence, PRICAI 2019, 26-30 August, 2019, Cuvu, Yanuka Island, Fiji , 2019. Abstract | BibTeX | Tags: Autonomous Systems | Links: @conference{@Prafull2019, title = {Optimizing DNN architectures for high speed autonomous navigation in GPS denied environments on edge devices}, author = {Prafull Prakash; Chaitanya Murti; Saketha Nath; Chiranjib, Bhattacharyya}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Optimizing-DNN-architectures-for-high-speed-autonomous-navigation-in-GPS-denied-environments-on-edge-devices.pdf}, doi = {10.1007/978-3-030-29911-8_36}, year = {2019}, date = {2019-08-23}, booktitle = {Pacific Rim International Conference on Artificial Intelligence, PRICAI 2019, 26-30 August, 2019, Cuvu, Yanuka Island, Fiji }, pages = {468-481}, abstract = {We address the challenge of high speed autonomous navigation of micro aerial vehicles (MAVs) using DNNs in GPS-denied environments with limited computational resources; specifically, we use the ODROID XU4 and the Raspberry Pi 3. The high computation costs of using DNNs for inference, particularly in the absence of powerful GPUs, necessitates negotiating a tradeoff between accuracy and inference. We address this tradeoff by employing sparsified neural networks. To obtain such architectures, we propose a novel algorithm to find sparse “sub networks” of existing pre trained models. Contrary to existing pruning-only strategies, our proposal includes a novel exploration step that efficiently searches for a different, but identically sparse, architecture with better generalization abilities. We derive learning theoretic bounds that reinforce our empirical findings that the optimized network achieves comparable generalization to the original network. We show that using our algorithm it is possible to discover models which, on average, have upto 19x fewer parameters than those obtained using existing state of the art pruning methods on autonomous navigation datasets, and achieve upto 6x improvements on inference time compared to existing state of the art shallow models on the ODROID XU4 and Raspberry Pi 3. Last, we demonstrate that our sparsified models can complete autonomous navigation missions with speeds upto 4 m/s using the ODROID XU4, which existing state of the art methods fail to do.}, keywords = {Autonomous Systems}, pubstate = {published}, tppubtype = {conference} } We address the challenge of high speed autonomous navigation of micro aerial vehicles (MAVs) using DNNs in GPS-denied environments with limited computational resources; specifically, we use the ODROID XU4 and the Raspberry Pi 3. The high computation costs of using DNNs for inference, particularly in the absence of powerful GPUs, necessitates negotiating a tradeoff between accuracy and inference. We address this tradeoff by employing sparsified neural networks. To obtain such architectures, we propose a novel algorithm to find sparse “sub networks” of existing pre trained models. Contrary to existing pruning-only strategies, our proposal includes a novel exploration step that efficiently searches for a different, but identically sparse, architecture with better generalization abilities. We derive learning theoretic bounds that reinforce our empirical findings that the optimized network achieves comparable generalization to the original network. We show that using our algorithm it is possible to discover models which, on average, have upto 19x fewer parameters than those obtained using existing state of the art pruning methods on autonomous navigation datasets, and achieve upto 6x improvements on inference time compared to existing state of the art shallow models on the ODROID XU4 and Raspberry Pi 3. Last, we demonstrate that our sparsified models can complete autonomous navigation missions with speeds upto 4 m/s using the ODROID XU4, which existing state of the art methods fail to do. |
| 34. | S. Kolathaya, Veer S PD based Robust Quadratic Programs for Robotic Systems Conference 29th IEEE International Conference on Intelligent Robots and Systems (IROS), 2019. Abstract | BibTeX | Tags: Robotics | Links: @conference{@Shishir2019, title = {PD based Robust Quadratic Programs for Robotic Systems}, author = {S. Kolathaya, S. Veer}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/12/PD-based-Robust-Quadratic-Programs-for-Robotic-Systems.pdf}, year = {2019}, date = {2019-11-04}, booktitle = {29th IEEE International Conference on Intelligent Robots and Systems (IROS)}, abstract = { In this paper, inspired by Proportional-Derivative(PD) control laws, we present a class of Control LyapunovFunction (CLF) based Quadratic Programs (QPs) for roboticsystems. Proportional-Derivative (PD) control laws are indepen-dent of the robot model, however, they fail to incorporate phys-ical constraints, such as torque saturation. On the other hand,most optimization based control design approaches ensuresatisfaction of the physical constraints, but they are sensitive toerrors in the robot model. The PD based Quadratic Programs(PD-QPs), presented in this paper, are a first step towardsbridging this gap between the PD and the optimization basedcontrollers to bring the best of both together. We derive twoversions of PD-QPs: model-based and model-free. Furthermore,for tracking time-varying trajectories, we establish asymptoticstability for the model-based PD-QP, and ultimate boundednessfor the model-free PD-QP. The performance of the PD-QPs isevaluated on two robot models: a fully actuated cart-pole andan underactuated5-DOF biped.}, keywords = {Robotics}, pubstate = {published}, tppubtype = {conference} } In this paper, inspired by Proportional-Derivative(PD) control laws, we present a class of Control LyapunovFunction (CLF) based Quadratic Programs (QPs) for roboticsystems. Proportional-Derivative (PD) control laws are indepen-dent of the robot model, however, they fail to incorporate phys-ical constraints, such as torque saturation. On the other hand,most optimization based control design approaches ensuresatisfaction of the physical constraints, but they are sensitive toerrors in the robot model. The PD based Quadratic Programs(PD-QPs), presented in this paper, are a first step towardsbridging this gap between the PD and the optimization basedcontrollers to bring the best of both together. We derive twoversions of PD-QPs: model-based and model-free. Furthermore,for tracking time-varying trajectories, we establish asymptoticstability for the model-based PD-QP, and ultimate boundednessfor the model-free PD-QP. The performance of the PD-QPs isevaluated on two robot models: a fully actuated cart-pole andan underactuated5-DOF biped. |
| 35. | Prasanth, V; Parekhji, Rubin; Amrutur, Bharadwaj Perturbation based workload augmentation for comprehensive functional safety analysis Conference Forthcoming Proceedings of the 32nd International Conference on VLSI Design and 18th International Conference on Embedded Systems, 05.-09.01.19, New Delhi, Forthcoming. BibTeX | Tags: Electrical Communications Engineering @conference{Prasanth2019, title = {Perturbation based workload augmentation for comprehensive functional safety analysis}, author = {V. Prasanth and Rubin Parekhji and Bharadwaj Amrutur}, year = {2019}, date = {2019-01-09}, booktitle = {Proceedings of the 32nd International Conference on VLSI Design and 18th International Conference on Embedded Systems, 05.-09.01.19, New Delhi}, keywords = {Electrical Communications Engineering}, pubstate = {forthcoming}, tppubtype = {conference} } |
| 36. | Abhishek Vijeev; Vinod Ganapathy; Chiranjib, Bhattacharyya Regulating Drones in Restricted Spaces Conference Proceedings of the 20th International Workshop on Mobile Computing Systems and Applications, Santa Cruz, CA, USA — February 27 - 28, 2019. Abstract | BibTeX | Tags: Drones | Links: @conference{@Abhishek2019, title = {Regulating Drones in Restricted Spaces}, author = {Abhishek Vijeev; Vinod Ganapathy; Chiranjib, Bhattacharyya}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Regulating-drones-in-restricted-spaces.pdf}, doi = {10.1145/3301293.3302370}, year = {2019}, date = {2019-02-22}, booktitle = {Proceedings of the 20th International Workshop on Mobile Computing Systems and Applications, Santa Cruz, CA, USA — February 27 - 28}, pages = {27-32}, abstract = {Commercial and end-user drones come equipped with a wide array of sensors. Unregulated use of such drones in public airspaces poses a serious threat to the privacy of citizens. We make the case for restricted spaces for drones, which are geographic areas for which a host can specify its privacy policies. Guest drones must prove to the host that they are in compliance with the host's policies before entering the restricted space. We then make the case for an information-flow control-based policy enforcement framework on drones, and sketch the design of a prototype framework atop the Robot Operating System (ROS).}, keywords = {Drones}, pubstate = {published}, tppubtype = {conference} } Commercial and end-user drones come equipped with a wide array of sensors. Unregulated use of such drones in public airspaces poses a serious threat to the privacy of citizens. We make the case for restricted spaces for drones, which are geographic areas for which a host can specify its privacy policies. Guest drones must prove to the host that they are in compliance with the host's policies before entering the restricted space. We then make the case for an information-flow control-based policy enforcement framework on drones, and sketch the design of a prototype framework atop the Robot Operating System (ROS). |
| 37. | Nihesh Rathod; Rajesh, Sundaresan Relay Placement Algorithms for Communication in a Heterogeneous Propagation Environment Conference 12th IFIP Wireless and Mobile Networking Conference September 11-13, 2019 Paris, France, 2019. Abstract | BibTeX | Tags: Electrical Communications Engineering | Links: @conference{@Nihesh2019, title = {Relay Placement Algorithms for Communication in a Heterogeneous Propagation Environment}, author = {Nihesh Rathod; Rajesh, Sundaresan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Relay-Placement-Algorithms-for-Communication-in-a-Heterogeneous-Propagation-Environment.pdf}, year = {2019}, date = {2019-09-11}, booktitle = {12th IFIP Wireless and Mobile Networking Conference September 11-13, 2019 Paris, France}, journal = {12th IFIP Wireless and Mobile Networking Conference September 11-13, 2019 Paris, France}, abstract = {A vast majority of the IoT devices will be connected in a topology where edge-devices push data to a local gateway which in turn forward the data to the cloud for analytics and archiving. In large networks, many edge-nodes will be located far away from their nearest gateways. Due to their locations and limited transmission power, these edge-nodes might experience poor network coverage. To provide reliable connectivity to these edge-nodes, relays/repeaters may have to be placed at a few locations until a gateway is reached. Network deployment engineers often do this based on field visits, surveys, measurements, initial deployments, followed by fine-tuning. For the scalability of such deployments, automated network planning tools are essential. Such tools should be able to predict coverage based on the edgenode locations using perhaps GIS data, identify the need for relays/repeater and, if needed, suggest the number and locations of relays, in order to provide reliable connectivity. This entire process should be as fast and automated as possible for rapid network deployment. In this paper, using a black box RSSI estimator between any candidate pair of transceiver locations, we propose a network deployment framework that uses either Ant Colony Optimisation (ACO) or Differential Evolution (DE) to identify the number and location of relays for meeting specified quality of service constraints. We also show how to handle multiple gateways.}, keywords = {Electrical Communications Engineering}, pubstate = {published}, tppubtype = {conference} } A vast majority of the IoT devices will be connected in a topology where edge-devices push data to a local gateway which in turn forward the data to the cloud for analytics and archiving. In large networks, many edge-nodes will be located far away from their nearest gateways. Due to their locations and limited transmission power, these edge-nodes might experience poor network coverage. To provide reliable connectivity to these edge-nodes, relays/repeaters may have to be placed at a few locations until a gateway is reached. Network deployment engineers often do this based on field visits, surveys, measurements, initial deployments, followed by fine-tuning. For the scalability of such deployments, automated network planning tools are essential. Such tools should be able to predict coverage based on the edgenode locations using perhaps GIS data, identify the need for relays/repeater and, if needed, suggest the number and locations of relays, in order to provide reliable connectivity. This entire process should be as fast and automated as possible for rapid network deployment. In this paper, using a black box RSSI estimator between any candidate pair of transceiver locations, we propose a network deployment framework that uses either Ant Colony Optimisation (ACO) or Differential Evolution (DE) to identify the number and location of relays for meeting specified quality of service constraints. We also show how to handle multiple gateways. |
| 38. | Alexandre Reiffers-Masson; Rajesh, Sundaresan Reputation-Based Information Design for Inducing Prosocial Behavior Journal Article arXiv:1905.00585 - Social and Information Networks, 2019. Abstract | BibTeX | Tags: Social and Information Networks | Links: @article{Alexandre2019, title = {Reputation-Based Information Design for Inducing Prosocial Behavior}, author = {Alexandre Reiffers-Masson; Rajesh, Sundaresan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Reputation-based-information-design-for-Inducing-Prosocial-Behavior.pdf}, year = {2019}, date = {2019-05-02}, journal = { arXiv:1905.00585 - Social and Information Networks}, abstract = {We study the idea of information design for inducing prosocial behavior in the context of electricity consumption. We consider a continuum of agents. Each agent has a different intrinsic motivation to reduce her power consumption. Each agent models the power consumption of the others via a distribution. Using this distribution, agents will anticipate their reputational benefit and choose a power consumption by trading off their own intrinsic motivation to do a prosocial action, the cost of this prosocial action and their reputation. Initially, the service provider can provide two types of quantized feedbacks of the power consumption. We study their advantages and disadvantages. For each feedback, we characterize the corresponding mean field equilibrium, using a fixed point equation. Besides computing the mean field equilibrium, we highlight the need for a systematic study of information design, by showing that revealing less information to the society can lead to more prosociality. In the last part of the paper, we introduce the notion of privacy and provide a new quantized feedback, more flexible than the previous ones, that respects agents' privacy concern but at the same time improves prosociality. The results of this study are also applicable to generic resource sharing problems.}, keywords = {Social and Information Networks}, pubstate = {published}, tppubtype = {article} } We study the idea of information design for inducing prosocial behavior in the context of electricity consumption. We consider a continuum of agents. Each agent has a different intrinsic motivation to reduce her power consumption. Each agent models the power consumption of the others via a distribution. Using this distribution, agents will anticipate their reputational benefit and choose a power consumption by trading off their own intrinsic motivation to do a prosocial action, the cost of this prosocial action and their reputation. Initially, the service provider can provide two types of quantized feedbacks of the power consumption. We study their advantages and disadvantages. For each feedback, we characterize the corresponding mean field equilibrium, using a fixed point equation. Besides computing the mean field equilibrium, we highlight the need for a systematic study of information design, by showing that revealing less information to the society can lead to more prosociality. In the last part of the paper, we introduce the notion of privacy and provide a new quantized feedback, more flexible than the previous ones, that respects agents' privacy concern but at the same time improves prosociality. The results of this study are also applicable to generic resource sharing problems. |
| 39. | Nayak, Ashika; Kundu, Atreyee Stabilizing switching signals for discrete-time switched linear systems: A unifying framework Conference Proceedings of the Fifth Indian Control Conference (ICC), 09.-11.01.19, New Delhi, 2019. BibTeX | Tags: Control and Optimisation @conference{Nayak2019, title = {Stabilizing switching signals for discrete-time switched linear systems: A unifying framework}, author = {Ashika Nayak and Atreyee Kundu}, year = {2019}, date = {2019-01-11}, booktitle = {Proceedings of the Fifth Indian Control Conference (ICC), 09.-11.01.19, New Delhi}, keywords = {Control and Optimisation}, pubstate = {published}, tppubtype = {conference} } |
| 40. | Ashwin K P; Ghosal, Ashitava Static Modeling of Miniaturized Pneumatic Artificial Muscles, Kinematic Analysis, and Experiments on an Endoscopic End-Effector Journal Article IEEE/ASME Transactions on Mechatronics, 24 (4), pp. 1429-1439, 2019. Abstract | BibTeX | Tags: Cyber Surgery and Remote Patient Care | Links: @article{Ashwin2019b, title = {Static Modeling of Miniaturized Pneumatic Artificial Muscles, Kinematic Analysis, and Experiments on an Endoscopic End-Effector}, author = {Ashwin, K P; Ghosal, Ashitava}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Static-Modeling-of-Miniaturized-Pneumatic-artificial-muscles.....pdf}, doi = {10.1109/TMECH.2019.2916783}, year = {2019}, date = {2019-05-14}, journal = { IEEE/ASME Transactions on Mechatronics}, volume = {24}, number = {4}, pages = {1429-1439}, abstract = {In this paper, we present the design, development, modeling, and experimental validation of an endoscopic attachment that can be used to independently position an endoscopic catheter tip at a desired location. Three miniaturized pneumatic artificial muscles (MPAMs) are used in a flexible endoscopic attachment, each MPAM is of 1.2 mm diameter and 45 mm in length and placed approximately 120 degrees apart within a pair of concentric springs. Pressurizing one or more MPAMs allows the tip to be positioned in a workspace, which is approximately a hemispherical section of radius 45 mm. We present a new and improved theoretical model for pressure-deformation relationship of a MPAM using static equations of a pressurized thick cylinder and constraints due to the braids. Comparison with existing models show that the proposed model performs better and the errors predicted by the model are less than 5% with experiments. A new forward kinematic model relating the position and orientation of the tip of the end-effector with changes in MPAM lengths is developed. Finally, we present experimental results conducted on a prototype endoscopic attachment and show that our model could predict the pose of the end-effector with a maximum error of 2 ± 1 mm.}, keywords = {Cyber Surgery and Remote Patient Care}, pubstate = {published}, tppubtype = {article} } In this paper, we present the design, development, modeling, and experimental validation of an endoscopic attachment that can be used to independently position an endoscopic catheter tip at a desired location. Three miniaturized pneumatic artificial muscles (MPAMs) are used in a flexible endoscopic attachment, each MPAM is of 1.2 mm diameter and 45 mm in length and placed approximately 120 degrees apart within a pair of concentric springs. Pressurizing one or more MPAMs allows the tip to be positioned in a workspace, which is approximately a hemispherical section of radius 45 mm. We present a new and improved theoretical model for pressure-deformation relationship of a MPAM using static equations of a pressurized thick cylinder and constraints due to the braids. Comparison with existing models show that the proposed model performs better and the errors predicted by the model are less than 5% with experiments. A new forward kinematic model relating the position and orientation of the tip of the end-effector with changes in MPAM lengths is developed. Finally, we present experimental results conducted on a prototype endoscopic attachment and show that our model could predict the pose of the end-effector with a maximum error of 2 ± 1 mm. |
| 41. | Mohammad Javad Khojasteh; Pavankumar Tallapragada; Jorge Cortés; Massimo, Franceschetti The value of timing information in event-triggered control Journal Article IEEE Transactions on Automatic Control, 2019. Abstract | BibTeX | Tags: Control and Optimisation, Networked Control systems | Links: @article{@Mohammad2019, title = {The value of timing information in event-triggered control}, author = {Mohammad Javad Khojasteh; Pavankumar Tallapragada; Jorge Cortés; Massimo,Franceschetti}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/The-value-of-timing-information-in-event-triggered-control-1.pdf}, doi = {10.1109/TAC.2019.2919107}, year = {2019}, date = {2019-05-27}, journal = {IEEE Transactions on Automatic Control}, abstract = {We study event-triggered control for stabilization of unstable linear plants over rate-limited communication channels subject to unknown, bounded delay. On one hand, the timing of event triggering carries implicit information about the state of the plant. On the other hand, the delay in the communication channel causes information loss, as it makes the state information available at the controller out of date. Combining these two effects, we show a phase transition behavior in the transmission rate required for stabilization using a given event-triggering strategy. For small values of the delay, the timing information carried by the triggering events is substantial, and the system can be stabilized with any positive rate. When the delay exceeds a critical threshold, the timing information alone is not enough to achieve stabilization, and the required rate grows. When the delay equals the inverse of the entropy rate of the plant, the implicit information carried by the triggering events perfectly compensates the loss of information due to the communication delay, and we recover the rate requirement prescribed by the data-rate theorem. We also provide an explicit construction yielding a sufficient rate for stabilization, as well as results for vector systems. Our results do not rely on any a priori probabilistic model for the delay or the initial conditions.}, keywords = {Control and Optimisation, Networked Control systems}, pubstate = {published}, tppubtype = {article} } We study event-triggered control for stabilization of unstable linear plants over rate-limited communication channels subject to unknown, bounded delay. On one hand, the timing of event triggering carries implicit information about the state of the plant. On the other hand, the delay in the communication channel causes information loss, as it makes the state information available at the controller out of date. Combining these two effects, we show a phase transition behavior in the transmission rate required for stabilization using a given event-triggering strategy. For small values of the delay, the timing information carried by the triggering events is substantial, and the system can be stabilized with any positive rate. When the delay exceeds a critical threshold, the timing information alone is not enough to achieve stabilization, and the required rate grows. When the delay equals the inverse of the entropy rate of the plant, the implicit information carried by the triggering events perfectly compensates the loss of information due to the communication delay, and we recover the rate requirement prescribed by the data-rate theorem. We also provide an explicit construction yielding a sufficient rate for stabilization, as well as results for vector systems. Our results do not rely on any a priori probabilistic model for the delay or the initial conditions. |
| 42. | Shishir Kolathaya; Dhaivat Dholakiya; Shalabh Bhatnagar; Abhik Singla; Shounak Bhattacharya; Ashitava Ghosal; Bharadwaj Amrutur; Abhimanyu Singh; Ashish Joglekar; Aditya Varma Sagi; Suhan Shetty; Ajay, Gunalan Trajectory Based Deep Policy Search for Quadrupedal Walking Conference The 28th IEEE International Conference on Robot & Human Interactive Communication, 2019. Abstract | BibTeX | Tags: Deep-RL, Quadruped, Robotics | Links: @conference{Kolathaya2019, title = {Trajectory Based Deep Policy Search for Quadrupedal Walking}, author = {Shishir Kolathaya; Dhaivat Dholakiya; Shalabh Bhatnagar; Abhik Singla; Shounak Bhattacharya; Ashitava Ghosal; Bharadwaj Amrutur; Abhimanyu Singh; Ashish Joglekar; Aditya Varma Sagi; Suhan Shetty; Ajay, Gunalan}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/Trajectory-based-Deep-policy-search-for-Quadrupedal-Walking.pdf}, year = {2019}, date = {2019-10-14}, booktitle = {The 28th IEEE International Conference on Robot & Human Interactive Communication}, abstract = {Abstract— In this paper, we explore a specific form of deep reinforcement learning (D-RL) technique for quadrupedal walking—trajectory based policy search via deep policy net-works. Existing approaches determine optimal policies for each time step, whereas we propose to determine an optimal policy for each walking step. We justify our approach based on the fact that animals including humans use “low” dimensional trajectories at the joint level to realize walking. We will construct these trajectories by using Bezier polynomials, with the coefficients being determined by a parameterized policy. In order to maintain smoothness of the trajectories during step transitions, hybrid invariance conditions are also applied. The policy action is computed at the beginning of every step, and a linear PD control law is applied to track at the individual joints. At the end of the step, reward is computed, which is then used to update the new policy parameters for the next step. Simulation results show that the number of training iterations are significantly reduced; approximately three times smaller in comparison with our previous methods for walking. After learning an optimal policy, i.e., an optimal walking gait for each step, we then successfully play them in a custom built quadruped robot, Stoch 2, thereby validating our approach.}, keywords = {Deep-RL, Quadruped, Robotics}, pubstate = {published}, tppubtype = {conference} } Abstract— In this paper, we explore a specific form of deep reinforcement learning (D-RL) technique for quadrupedal walking—trajectory based policy search via deep policy net-works. Existing approaches determine optimal policies for each time step, whereas we propose to determine an optimal policy for each walking step. We justify our approach based on the fact that animals including humans use “low” dimensional trajectories at the joint level to realize walking. We will construct these trajectories by using Bezier polynomials, with the coefficients being determined by a parameterized policy. In order to maintain smoothness of the trajectories during step transitions, hybrid invariance conditions are also applied. The policy action is computed at the beginning of every step, and a linear PD control law is applied to track at the individual joints. At the end of the step, reward is computed, which is then used to update the new policy parameters for the next step. Simulation results show that the number of training iterations are significantly reduced; approximately three times smaller in comparison with our previous methods for walking. After learning an optimal policy, i.e., an optimal walking gait for each step, we then successfully play them in a custom built quadruped robot, Stoch 2, thereby validating our approach. |
| 43. | Jogendra Nath Kundu; Nishank Lakkakula; R, Venkatesh Babu UM-Adapt: Unsupervised Multi-Task Adaptation Using Adversarial Cross-Task Distillation Journal Article arXiv:1908.03884, Computer Vision and Pattern Recognition, 2019. Abstract | BibTeX | Tags: Student Research Grant | Links: @article{Jogendra2019, title = {UM-Adapt: Unsupervised Multi-Task Adaptation Using Adversarial Cross-Task Distillation}, author = {Jogendra Nath Kundu; Nishank Lakkakula; R, Venkatesh Babu}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/10/UM-Adapt-Unsupervised-multi-task-adaptation-using-Adversarial-cross-task-distillation.pdf}, year = {2019}, date = {2019-09-16}, journal = {arXiv:1908.03884, Computer Vision and Pattern Recognition}, abstract = {Aiming towards human-level generalization, there is a need to explore adaptable representation learning methods with greater transferability. Most existing approaches independently address task-transferability and cross-domain adaptation, resulting in limited generalization. In this paper, we propose UM-Adapt - a unified framework to effectively perform unsupervised domain adaptation for spatially-structured prediction tasks, simultaneously maintaining a balanced performance across individual tasks in a multi-task setting. To realize this, we propose two novel regularization strategies; a) Contour-based content regularization (CCR) and b) exploitation of inter-task coherency using a cross-task distillation module. Furthermore, avoiding a conventional ad-hoc domain discriminator, we re-utilize the cross-task distillation loss as output of an energy function to adversarially minimize the input domain discrepancy. Through extensive experiments, we demonstrate superior generalizability of the learned representations simultaneously for multiple tasks under domain-shifts from synthetic to natural environments. UM-Adapt yields state-of-the-art transfer learning results on ImageNet classification and comparable performance on PASCAL VOC 2007 detection task, even with a smaller backbone-net. Moreover, the resulting semi-supervised framework outperforms the current fully-supervised multi-task learning state-of-the-art on both NYUD and Cityscapes dataset. }, keywords = {Student Research Grant}, pubstate = {published}, tppubtype = {article} } Aiming towards human-level generalization, there is a need to explore adaptable representation learning methods with greater transferability. Most existing approaches independently address task-transferability and cross-domain adaptation, resulting in limited generalization. In this paper, we propose UM-Adapt - a unified framework to effectively perform unsupervised domain adaptation for spatially-structured prediction tasks, simultaneously maintaining a balanced performance across individual tasks in a multi-task setting. To realize this, we propose two novel regularization strategies; a) Contour-based content regularization (CCR) and b) exploitation of inter-task coherency using a cross-task distillation module. Furthermore, avoiding a conventional ad-hoc domain discriminator, we re-utilize the cross-task distillation loss as output of an energy function to adversarially minimize the input domain discrepancy. Through extensive experiments, we demonstrate superior generalizability of the learned representations simultaneously for multiple tasks under domain-shifts from synthetic to natural environments. UM-Adapt yields state-of-the-art transfer learning results on ImageNet classification and comparable performance on PASCAL VOC 2007 detection task, even with a smaller backbone-net. Moreover, the resulting semi-supervised framework outperforms the current fully-supervised multi-task learning state-of-the-art on both NYUD and Cityscapes dataset. |
| 44. | Poorna Chandra, Tejasvi User story contest @ RabbitMQ Summit Presentation 04.11.2019. Abstract | BibTeX | Tags: Smart City @misc{@Poorna2019, title = {User story contest @ RabbitMQ Summit}, author = {Poorna Chandra, Tejasvi }, year = {2019}, date = {2019-11-04}, abstract = {RabbitMQ Summit brings together users and developers from around the world. Learn from speakers and keynotes on what’s happening in and around RabbitMQ, and how top companies utilise RabbitMQ to power their services. }, keywords = {Smart City}, pubstate = {published}, tppubtype = {presentation} } RabbitMQ Summit brings together users and developers from around the world. Learn from speakers and keynotes on what’s happening in and around RabbitMQ, and how top companies utilise RabbitMQ to power their services. |
2018 |
|
| 45. | Ramaswamy, Arunselvan; Bhatnagar, Shalabh Analysis of gradient descent methods with nondiminishing bounded errors Journal Article IEEE Transactions on Automatic Control, 63 (5), pp. 1465-1471, 2018. Abstract | BibTeX | Tags: Control and Optimisation | Links: @article{Ramaswamy2018, title = { Analysis of gradient descent methods with nondiminishing bounded errors}, author = {Arunselvan Ramaswamy and Shalabh Bhatnagar}, url = {http://www.rbccps.org/wp-content/uploads/2018/08/08016343.pdf}, doi = {10.1109/TAC.2017.2744598}, year = {2018}, date = {2018-05-31}, journal = {IEEE Transactions on Automatic Control}, volume = {63}, number = {5}, pages = {1465-1471}, abstract = {The main aim of this paper is to provide an analysis of gradient descent (GD) algorithms with gradient errors that do not necessarily vanish, asymptotically. In particular, sufficient conditions are presented for both stability (almost sure boundedness of the iterates) and convergence of GD with bounded (possibly) nondiminishing gradient errors. In addition to ensuring stability, such an algorithm is shown to converge to a small neighborhood of the minimum set, which depends on the gradient errors. It is worth noting that the main result of this paper can be used to show that GD with asymptotically vanishing errors indeed converges to the minimum set. The results presented herein are not only more general when compared to previous results, but our analysis of GD with errors is new to the literature to the best of our knowledge. Our work extends the contributions of Mangasarian and Solodov, Bertsekas and Tsitsiklis, and Tadiόand Doucet. Using our framework, a simple yet effective implementation of GD using simultaneous perturbation stochastic approximations, with constant sensitivity parameters, is presented. Another important improvement over many previous results is that there are no “additional” restrictions imposed on the step sizes. In machine learning applications where step sizes are related to learning rates, our assumptions, unlike those of other papers, do not affect these learning rates. Finally, we present experimental results to validate our theory.}, keywords = {Control and Optimisation}, pubstate = {published}, tppubtype = {article} } The main aim of this paper is to provide an analysis of gradient descent (GD) algorithms with gradient errors that do not necessarily vanish, asymptotically. In particular, sufficient conditions are presented for both stability (almost sure boundedness of the iterates) and convergence of GD with bounded (possibly) nondiminishing gradient errors. In addition to ensuring stability, such an algorithm is shown to converge to a small neighborhood of the minimum set, which depends on the gradient errors. It is worth noting that the main result of this paper can be used to show that GD with asymptotically vanishing errors indeed converges to the minimum set. The results presented herein are not only more general when compared to previous results, but our analysis of GD with errors is new to the literature to the best of our knowledge. Our work extends the contributions of Mangasarian and Solodov, Bertsekas and Tsitsiklis, and Tadiόand Doucet. Using our framework, a simple yet effective implementation of GD using simultaneous perturbation stochastic approximations, with constant sensitivity parameters, is presented. Another important improvement over many previous results is that there are no “additional” restrictions imposed on the step sizes. In machine learning applications where step sizes are related to learning rates, our assumptions, unlike those of other papers, do not affect these learning rates. Finally, we present experimental results to validate our theory. |
| 46. | Krishnapuram, Raghu; Rajkumar, Arun; Acharya, Adithya; Dhara, Nikhil; Goudar, Manjunath; Sarashetti, Akshay P Online domain adaptation by exploiting labeled features and pro-active learning Conference Proceedings of the 2018 ACM India Joint International Conference on Data Science and Management of Data (CoDS-COMAD), 11.-13.01.18, Goa, 2018. Abstract | BibTeX | Tags: Machine Learning | Links: @conference{Krishnapuram2018, title = { Online domain adaptation by exploiting labeled features and pro-active learning}, author = {Raghu Krishnapuram and Arun Rajkumar and Adithya Acharya and Nikhil Dhara and Manjunath Goudar and Akshay P. Sarashetti}, url = {http://www.rbccps.org/wp-content/uploads/2018/08/p137-krishnapuram.pdf}, doi = {10.1145/3152494.3152507}, year = {2018}, date = {2018-01-13}, booktitle = {Proceedings of the 2018 ACM India Joint International Conference on Data Science and Management of Data (CoDS-COMAD), 11.-13.01.18, Goa}, pages = {137-145}, abstract = {Domain adaptation algorithms have gained considerable interest in recent years due to their empirical success. However, most of the existing literature assumes the availability of a pool of unlabeled target domain data which is not the case in several practical scenarios. We consider the problem of online domain adaptation where the target (test) data arrives one at a time, and there is a fixed budget available for obtaining their labels. We propose a novel online domain adaptation algorithm which uses the budget judiciously by balancing the cost and reliability of the oracles which provide the labels. The proposed algorithm uses a probabilistic model which allows the seamless integration of previously unseen features (i.e., features that may appear during test time) into the model. Experiments on several benchmark real world datasets empirically establish the efficacy of the proposed algorithm.}, keywords = {Machine Learning}, pubstate = {published}, tppubtype = {conference} } Domain adaptation algorithms have gained considerable interest in recent years due to their empirical success. However, most of the existing literature assumes the availability of a pool of unlabeled target domain data which is not the case in several practical scenarios. We consider the problem of online domain adaptation where the target (test) data arrives one at a time, and there is a fixed budget available for obtaining their labels. We propose a novel online domain adaptation algorithm which uses the budget judiciously by balancing the cost and reliability of the oracles which provide the labels. The proposed algorithm uses a probabilistic model which allows the seamless integration of previously unseen features (i.e., features that may appear during test time) into the model. Experiments on several benchmark real world datasets empirically establish the efficacy of the proposed algorithm. |
| 47. | Naveen, K P; Sundaresan, Rajesh A double-auction mechanism for mobile data-offloading markets with strategic agents Conference Proceedings of the 16th International Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt), 07.-11.05.18, Shanghai (China), 2018. Abstract | BibTeX | Tags: Electrical Communications Engineering | Links: @conference{Naveen2018, title = {A double-auction mechanism for mobile data-offloading markets with strategic agents}, author = {K. P. Naveen and Rajesh Sundaresan}, url = {http://www.rbccps.org/wp-content/uploads/2018/12/08362857.pdf}, doi = {10.23919/WIOPT.2018.8362857}, year = {2018}, date = {2018-05-24}, booktitle = {Proceedings of the 16th International Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt), 07.-11.05.18, Shanghai (China)}, abstract = {We consider a recently proposed double-auction mechanism for mobile data-offloading. Network operators (users) derive benefit from offloading their traffic to third party WiFi or femtocell network (link-supplier). A link-supplier experiences costs for the additional capacity that he provides. Users and link-supplier (collectively referred to as agents) have their utilities and cost function as private knowledge. A system-designer decomposes the problem into a network problem (with surrogate utilities and surrogate cost functions) and agent problems (one per agent). The surrogate utilities and cost functions are modulated by the agents' bids. Agents' payoffs and costs are then determined by the allocations and prices set by the system designer. So long as the agents do not anticipate the effect of their actions, a competitive equilibrium exists as a solution to the network and agent problems, and this equilibrium optimizes the system utility. This work shows that when the agents are strategic (price-anticipating), the presence of strategic supplying agents drives the system to an undesirable equilibrium with zero participation. This is in stark contrast to the setting when link-suppliers are not strategic where the efficiency loss is at most 34%. The paper then proposes a Stackelberg game modification to alleviate the efficiency loss problem. The system designer first announces the allocation and payment functions. He then invites the supplying agents to announce their bids. He then invites the users to respond to the suppliers' bids. The resulting efficiency loss is characterized in terms of the suppliers' cost functions.}, keywords = {Electrical Communications Engineering}, pubstate = {published}, tppubtype = {conference} } We consider a recently proposed double-auction mechanism for mobile data-offloading. Network operators (users) derive benefit from offloading their traffic to third party WiFi or femtocell network (link-supplier). A link-supplier experiences costs for the additional capacity that he provides. Users and link-supplier (collectively referred to as agents) have their utilities and cost function as private knowledge. A system-designer decomposes the problem into a network problem (with surrogate utilities and surrogate cost functions) and agent problems (one per agent). The surrogate utilities and cost functions are modulated by the agents' bids. Agents' payoffs and costs are then determined by the allocations and prices set by the system designer. So long as the agents do not anticipate the effect of their actions, a competitive equilibrium exists as a solution to the network and agent problems, and this equilibrium optimizes the system utility. This work shows that when the agents are strategic (price-anticipating), the presence of strategic supplying agents drives the system to an undesirable equilibrium with zero participation. This is in stark contrast to the setting when link-suppliers are not strategic where the efficiency loss is at most 34%. The paper then proposes a Stackelberg game modification to alleviate the efficiency loss problem. The system designer first announces the allocation and payment functions. He then invites the supplying agents to announce their bids. He then invites the users to respond to the suppliers' bids. The resulting efficiency loss is characterized in terms of the suppliers' cost functions. |
| 48. | Alexandre Reiffers Masson; Yezekael Hayel; Eitan Altman; Guillaume, Martel A Generalized Fractional Program for Maximizing Content Popularity in Online Social Networks Conference 2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), Barcelona, Spain, 2018. Abstract | BibTeX | Tags: Social and Information Networks | Links: @conference{@Alex2018c, title = {A Generalized Fractional Program for Maximizing Content Popularity in Online Social Networks}, author = {Alexandre Reiffers Masson; Yezekael Hayel; Eitan Altman; Guillaume, Martel}, url = {http://cps.iisc.ac.in/wp-content/uploads/2019/11/A-Generalized-Fractional-Program-for-Maximizing-Content-Popularity-in-Online-Social-Networks.pdf}, doi = {10.1109/ASONAM.2018.8508480}, year = {2018}, date = {2018-08-01}, booktitle = {2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), Barcelona, Spain}, pages = {869-872}, abstract = {In this paper, we consider a ”generalized” fractional program in order to solve a popularity optimization problem in which a source of contents controls the topics of her contents and the rate with which posts are sent to a time line. The objective of the source is to maximize its overall popularity in an Online Social Network (OSN). We propose an efficient algorithm that converges to the optimal solution of the Popularity maximization problem.}, keywords = {Social and Information Networks}, pubstate = {published}, tppubtype = {conference} } In this paper, we consider a ”generalized” fractional program in order to solve a popularity optimization problem in which a source of contents controls the topics of her contents and the rate with which posts are sent to a time line. The objective of the source is to maximize its overall popularity in an Online Social Network (OSN). We propose an efficient algorithm that converges to the optimal solution of the Popularity maximization problem. |
| 49. | Lakshminarayanan, Chandrashekar; Bhatnagar, Shalabh; Szepesvári, Csaba A linearly relaxed approximate linear program for Markov decision processes Journal Article IEEE Transactions on Automatic Control, 63 (4), pp. 1185-1191, 2018. Abstract | BibTeX | Tags: Control and Optimisation | Links: @article{Lakshminarayanan2018, title = {A linearly relaxed approximate linear program for Markov decision processes}, author = {Chandrashekar Lakshminarayanan and Shalabh Bhatnagar and Csaba Szepesvári}, url = {http://www.rbccps.org/wp-content/uploads/2018/08/08014469.pdf}, doi = {10.1109/TAC.2017.2743163}, year = {2018}, date = {2018-04-30}, journal = {IEEE Transactions on Automatic Control}, volume = {63}, number = {4}, pages = {1185-1191}, abstract = {Approximate linear programming (ALP) and its variants have been widely applied to Markov decision processes (MDPs) with a large number of states. A serious limitation of ALP is that it has an intractable number of constraints, as a result of which constraint approximations are of interest. In this paper, we define a linearly relaxed approximation linear program (LRALP) that has a tractable number of constraints, obtained as positive linear combinations of the original constraints of the ALP. The main contribution is a novel performance bound for LRALP.}, keywords = {Control and Optimisation}, pubstate = {published}, tppubtype = {article} } Approximate linear programming (ALP) and its variants have been widely applied to Markov decision processes (MDPs) with a large number of states. A serious limitation of ALP is that it has an intractable number of constraints, as a result of which constraint approximations are of interest. In this paper, we define a linearly relaxed approximation linear program (LRALP) that has a tractable number of constraints, obtained as positive linear combinations of the original constraints of the ALP. The main contribution is a novel performance bound for LRALP. |
| 50. | Raseena, K T; Ghosh, Prasanta Kumar A maximum likelihood formulation to exploit heart rate variability for robust heart rate estimation from facial video Conference Proceedings of the 40th International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 18.-21.07., Honolulu (USA), 2018. Abstract | BibTeX | Tags: Student Research Grant | Links: @conference{Raseena2018, title = {A maximum likelihood formulation to exploit heart rate variability for robust heart rate estimation from facial video}, author = {K. T. Raseena and Prasanta Kumar Ghosh}, url = {http://www.rbccps.org/wp-content/uploads/2018/12/08513483.pdf}, doi = {10.1109/EMBC.2018.8513483}, year = {2018}, date = {2018-10-29}, booktitle = {Proceedings of the 40th International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 18.-21.07., Honolulu (USA)}, abstract = {The problem of estimating the heart rate (HR) from a facial video is considered. A typical approach for this problem is to use independent component analysis (ICA) on the red, blue, green intensity profiles averaged over the facial region. This provides estimates of the underlying source signals, whose spectral peaks are used to predict HR in every analysis window. In this work, we propose a maximum likelihood formulation to optimally select a source signal in each window such that the predicted HR trajectory not only corresponds to the most likely spectral peaks but also ensures a realistic HR variability (HRV) across analysis windows. The likelihood function is efficiently optimized using dynamic programming in a manner similar to Viterbi decoding. The proposed scheme for HR estimation is denoted by vICA. The performance of vICA is compared with a typical ICA approach as well as a recently proposed sparse spectral peak tracking (SSPT) technique that ensures that the predicted HR does not vary drastically across analysis windows. Experiments are performed in a five fold setup using facial videos of 15 subjects recorded using two types of smartphones (Samsung Galaxy and iPhone) at three different distances (6inches, 1foot, 2feet) between the phone camera and the subject. Mean absolute error (MAE) between the original and predicted HR reveals that the proposed vICA scheme performs better than the best of the baseline schemes, namely SSPT by -8.69%, 52.77% and 8.00% when Samsung Galaxy phone was used at a distance of 6inches, 1foot, and 2feet respectively. These improvements are 12.13%, 13.59% and 18.34% when iPhone was used. This, in turn, suggests that the HR predicted from a facial video becomes more accurate when the smoothness of HRV is utilized in predicting the HR trajectory as done in the proposed vICA.}, keywords = {Student Research Grant}, pubstate = {published}, tppubtype = {conference} } The problem of estimating the heart rate (HR) from a facial video is considered. A typical approach for this problem is to use independent component analysis (ICA) on the red, blue, green intensity profiles averaged over the facial region. This provides estimates of the underlying source signals, whose spectral peaks are used to predict HR in every analysis window. In this work, we propose a maximum likelihood formulation to optimally select a source signal in each window such that the predicted HR trajectory not only corresponds to the most likely spectral peaks but also ensures a realistic HR variability (HRV) across analysis windows. The likelihood function is efficiently optimized using dynamic programming in a manner similar to Viterbi decoding. The proposed scheme for HR estimation is denoted by vICA. The performance of vICA is compared with a typical ICA approach as well as a recently proposed sparse spectral peak tracking (SSPT) technique that ensures that the predicted HR does not vary drastically across analysis windows. Experiments are performed in a five fold setup using facial videos of 15 subjects recorded using two types of smartphones (Samsung Galaxy and iPhone) at three different distances (6inches, 1foot, 2feet) between the phone camera and the subject. Mean absolute error (MAE) between the original and predicted HR reveals that the proposed vICA scheme performs better than the best of the baseline schemes, namely SSPT by -8.69%, 52.77% and 8.00% when Samsung Galaxy phone was used at a distance of 6inches, 1foot, and 2feet respectively. These improvements are 12.13%, 13.59% and 18.34% when iPhone was used. This, in turn, suggests that the HR predicted from a facial video becomes more accurate when the smoothness of HRV is utilized in predicting the HR trajectory as done in the proposed vICA. |
