Siddharth Karamcheti
Institute of Occupational Medicine, Stanford University, Brown University
Papers
18
Total Citations
2,512
H-Index
8
About
Siddharth Karamcheti is a researcher at the intersection of natural language processing, robot learning, and human-robot interaction, with a focus on building intelligent systems that can understand and respond to human language in real-world settings. His work spans foundation models, visual representation learning, and language-guided robotics, addressing fundamental questions about how machines can learn from and adapt to human instruction. Karamcheti contributed to the landmark "Foundation Models" report (2021, 2,177 citations), one of the most influential AI papers of the decade, helping define the paradigm around large-scale pretrained models like GPT-3 and DALL-E. His robotics research has pushed the boundaries of language-driven learning, with notable contributions including DROID, a large-scale robot manipulation dataset (108 citations), and OpenVLA, an open-source vision-language-action model enabling more accessible robot policy learning. His work on adaptive natural language interfaces — teaching robots through decomposition and interactive feedback — reflects a consistent commitment to making robots genuinely responsive to human communication. More recently, he has explored grounded commonsense reasoning, asking how robots can make contextually appropriate decisions beyond literal instruction-following, a critical step toward trustworthy real-world deployment.
Research Focus
Key Achievements
Top Papers
- 1On the Opportunities and Risks of Foundation Models2,177 citations · 2021
- 2DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset108 citations · 2024
- 3Language-Driven Representation Learning for Robotics55 citations · 2023
- 4No, to the Right45 citations · 2023
- 5OpenVLA: An Open-Source Vision-Language-Action Model39 citations · 2024
- 6Learning Adaptive Language Interfaces through Decomposition22 citations · 2020
- 7
- 8Toward Grounded Commonsense Reasoning12 citations · 2024
- 9Learning Visually Guided Latent Actions for Assistive Teleoperation7 citations · 2021
- 10