Beyond Uniform Sampling: Offline Reinforcement Learning with Imbalanced Datasets
- Zhang-wei Hong
- Aviral Kumar
- et al.
- 2023
- NeurIPS 2023
Abhishek Bhandwaldar
Title
Research Software Engineer
Bio
Abhishek Bhandwaldar is a research software engineer in MIT-IBM Watson AI Lab. He has worked on developing a framework for conversational agents and business process automation and creating interactive game for educating the general public about machine learning. He currently works on creating 3D scenarios aimed at learning world model, and reinforcement learning. He earned his master’s in computer science from the University of North Carolina, Charlotte and his master’s thesis was on biomedical question and answering. Bhandwaldar received his undergraduate degree from Pune University.
Publications
The ThreeDWorld Transport Challenge: A Visually Guided Task-and-Motion Planning Benchmark Towards Physically Realistic Embodied AI
- Chuang Gan
- Siyuan Zhou
- et al.
- 2022
- ICRA 2022
ThreeDWorld: A Platform for Interactive Multi-Modal Physical Simulation
- Chuang Gan
- Jeremy Schwartz
- et al.
- 2021
- NeurIPS 2021
OPEn: An Open-ended Physics Environment for Learning Without a Task
- Chuang Gan
- Abhishek Bhandwaldar
- et al.
- 2021
- IROS 2021
AGENT: A Benchmark for Core Psychological Reasoning
- Tianmin Shu
- Abhishek Bhandwaldar
- et al.
- 2021
- ICML 2021
Blog posts
Evaluating common sense in AI
Deep DiveAbhishek Bhandwaldar and Tianmin Shu15 minute readIBM, MIT and Harvard release “Common Sense AI” dataset at ICML 2021
ReleaseDan Gutfreund, Abhishek Bhandwaldar, and Chuang Gan6 minute readMIT and IBM announce ThreeDWorld Transport Challenge for physically realistic Embodied AI
ReleaseChuang Gan, Abhishek Bhandwaldar, and Dan Gutfreund7 minute read
Top collaborators
DG
Dan Gutfreund
Principal Research Staff Member, Manager, AI Models Science
DC
David D. Cox
VP, AI Models; IBM Director, MIT-IBM Watson AI Lab
AS
Akash Srivastava
Director and Chief Architect, Core AI, IBM Software