Papers
22
Total Citations
3,384
H-Index
15
About
Abhishek Gupta is a leading researcher at the intersection of deep reinforcement learning (RL) and robotics, with a focus on enabling autonomous systems to acquire complex, real-world skills with high efficiency. His seminal work on **Soft Actor-Critic (SAC)** algorithms, cited over 1,950 times, revolutionized model-free RL by addressing the critical challenges of sample complexity and hyperparameter sensitivity, establishing a foundational framework for countless subsequent applications. Gupta has made major contributions to **dexterous manipulation**, demonstrating how deep RL can control both low-cost soft robotic hands and multi-fingered hands to perform intricate tasks, often by integrating human demonstrations to accelerate learning. His research on **imitation from observation** pioneered methods for agents to learn behaviors by watching raw video, bypassing the need for action labels and enabling skill transfer across different morphologies. A key collaborator on the **Open X-Embodiment** project, he is helping to consolidate diverse robotic datasets into general-purpose models. With over 3,000 total citations, Gupta’s work is shaping the future of sample-efficient, generalizable robotic learning.
Research Focus
Key Achievements
Top Papers
- 1Soft Actor-Critic Algorithms and Applications1,952 citations · 2018
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10