Reinforcement Learning

Learning algorithms that enable software agents to take actions that maximizes some notion of cumulative reward.

About

Inspired from behavioural psychology, reinforcement learning agents determine the ideal behaviour in a given context with the help of feedback. Our group focuses on both the theoretical and practical aspects of RL, more recently in the context of deep learning, with ongoing projects ranging from establishing theoretical bounds on augmented multi-armed bandit algorithms to achieving safe and stable learning algorithm for deployment in safety-critical applications like Autonomous Driving.

People

M.Tech and Dual Degree

Project Associates

Publications

  1. Tracking and Stabilization of Mechanical Systems using Reinforcement Learning B. Ravindran R. Pasumarthy S. Bhuvaneswari, A Mahindrakar To appear in the Proceedings of the Fourth Indian Control Conference (2018)
  2. An extended reinforcement learning model of basal ganglia to understand the contributions of serotonin and dopamine in risk-based decision making, reward prediction, and punishment learning Pragathi P Balasubramani, V Srinivasa Chakravarthy, Balaraman Ravindran, Ahmed A Moustafa Frontiers in computational neuroscience (2014) 8
  3. Learning to repeat: Fine grained action repetition for deep reinforcement learning Sahil Sharma, Aravind S Lakshminarayanan, Balaraman Ravindran arXiv preprint arXiv:1702.06054 (2017)
  4. EPOpt: Learning Robust Neural Network Policies Using Model Ensembles Aravind Rajeswaran, Sarvjeet Ghotra, Sergey Levine, Balaraman Ravindran arXiv preprint arXiv:1610.01283 (2016)
  5. Data Driven Strategies for Active Monocular SLAM using Inverse Reinforcement Learning Vignesh Prasad, Rishabh Jangir, Ravindran Balaraman, K Madhava Krishna Proceedings of the 16th Conference on Autonomous Agents and MultiAgent Systems (2017) 1697--1699
  6. Thresholding Bandits with Augmented UCB Subhojyoti Mukherjee, KP Naveen, Nandan Sudarsanam, Balaraman Ravindran arXiv preprint arXiv:1704.02281 (2017)
  7. $ A\\^{} 2T $: Attend, Adapt and Transfer: Attentive Deep Architecture for Adaptive Transfer from multiple sources Janarthanan Rajendran, Aravind Lakshminarayanan, Mitesh M Khapra, Balaraman Ravindran, others arXiv preprint arXiv:1510.02879 (2015)
  8. Dynamic Action Repetition for Deep Reinforcement Learning. Aravind S Lakshminarayanan, Sahil Sharma, Balaraman Ravindran AAAI (2017) 2133--2139
  9. Dynamic frame skip deep q network Aravind S Lakshminarayanan, Sahil Sharma, Balaraman Ravindran arXiv preprint arXiv:1605.05365 (2016)
  10. Activity recognition for natural human robot interaction Addwiteey Chrungoo, SS Manimaran, Balaraman Ravindran International Conference on Social Robotics (2014) 84--94
  11. RRTPI: Policy iteration on continuous domains using rapidly-exploring random trees Manimaran Sivasamy Sivamurugan, Balaraman Ravindran Robotics and Automation (ICRA), 2014 IEEE International Conference on (2014) 4362--4367