Home Research Robots Projects

Sriyash Poddar
poddarsriyash [at] iitkgp [dot] ac [dot] in

I am a final year undergraduate student at the Department of Computer Science and Engineering at IIT Kharagpur. My research interests lie at the intersection of human robot interaction and robot learning.

Currently, I work with Prof. Siddhartha Srinavasa and Christoforos Mavrogiannis at the University of Washington, on social navigation. My work is mainly on ensuring human safety and comfort in crowd navigation scenarios. I am also exploring how does human motion prediction affect robot control, grounding my work with real-world trials on the Honda P.A.T.H Bot.

During Summer 2022, I had the pleasure of working with Prof. Sarath Chandar at Mila - Quebec AI Insitute on dealing with non-stationarity in multi-agent learning. Previously, I also spend a summer at Adobe Research, working at the intersection of topological representation and lifelong learning.

I will be applying for PhD positions for the Fall 2023 admit.

email   |   cv   |   github   |   google scholar

profile photo
News

Publications

* indicates equal contribution

Winding Through: Crowd Navigation via Topological Invariance
Christoforos Mavrogiannis, Krishna Balasubramanian, Sriyash Poddar, Anush Gandra, Siddhartha S. Srinivasa
In IEEE Robotics and Automation Letters (RA-L) November 2022.

paper   video

TRACE: Topologically Constrained Reward Aware Action Embeddings for Life-Long Reinforcement Learning
Sriyash Poddar, Tanay Anand, Pinkesh Badjatiya, Jayakumar Subramanian, Georgios Theocharous, Balaji K.
Under review at NeurIPS Workshop on Human in the Loop Learning 2022.

paper

Understanding the Role of Affect Dimensions in Detecting Emotions from Tweets: A Multi-task Approach
Rajdeep Mukherjee, Sriyash Poddar*, Atharva Naik*, Soham Dasgupta, Niloy Ganguly
In the International ACM SIGIR Conference on Research and Development in Information Retrieval 2021.

paper   code   video

Experience


Chandar Lab, Mila-Quebec AI Institute
Advisor: Prof. Sarath Chandar

Topics: Multi-Agent RL, Intrinsic Motivation, Cooperative Games
Also contributing and maintaining RLHive : a univeral RL framework for singular and multi-agent training.

Personal Robotics Lab, University of Washington
Advisor: Prof. Siddhartha Srinavasa and Christoforos Mavrogiannis

Topics: Social Navigation, Human-Robot Interaction, Model Predictive Control.
Tested controllers for crowd navigation in real world trials with humans on the Honda P.A.T.H Bot.

Adobe Media and Data Science Research
Advisors: Jayakumar Subramaniam and Georgios Theocharous

Topics: Continual Learning, Topological Representations

Kharagpur Robosoccer Students Group

Topics: Humanoid Robots, Inverse Kinematic Control, Evolutionary Learning

Services

Teaching Assistant: Reinforcement Learning (CS60027)

IIT Kharagpur, Fall 2022
Mentoring students on the course projects and the ML Reproducibility Challenge 2022.
Instructor for lectures on multi-agent reinforcement learning.

Reading Group: Reinforcement Learning and Robotics

Organizer, Fall, Spring 2020

IEEE Winter School on Image Processing

IIT Kharagpur, Fall 2019
Mentored students at a two week bootcamp on image processing and machine learning.



website adapted from here