Papers
22
Total Citations
2,459
H-Index
15
About
Igor Mordatch is a pioneering researcher at the intersection of robotics, machine learning, and embodied artificial intelligence, whose work has fundamentally shaped how machines learn to understand and interact with the physical world. Based primarily at Google DeepMind, Mordatch has been instrumental in developing large-scale robotic learning systems, most notably contributing to the landmark RT-1, RT-2, and PaLM-E projects, which collectively demonstrate how transformer-based and vision-language models trained on internet-scale data can be transferred to real-world robotic control — work that has garnered over 1,100 citations combined. His influential research on "Inner Monologue" (206 citations) showed how language model reasoning could be embedded into embodied planning loops, advancing the frontier of semantic robot reasoning. Earlier contributions include pioneering simulation-to-real transfer techniques and dynamic humanoid motion planning, reflecting a career-long commitment to bridging the gap between theoretical models and physical systems. His involvement in Open X-Embodiment further underscores his dedication to open, collaborative robotics research. With a citation portfolio exceeding 2,100 across a decade of work, Mordatch stands as a defining voice in the era of foundation models for robotics.
Research Focus
Key Achievements
Top Papers
- 1RT-1: Robotics Transformer for Real-World Control at Scale512 citations · 2023
- 2PaLM-E: An Embodied Multimodal Language Model350 citations · 2023
- 3RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control267 citations · 2023
- 4Inner Monologue: Embodied Reasoning through Planning with Language Models206 citations · 2022
- 5Implicit Generation and Modeling in Energy-Based Models168 citations · 2019
- 6
- 7
- 8
- 9Implicit Generation and Generalization in Energy-Based Models112 citations · 2019
- 10Open X-Embodiment: Robotic Learning Datasets and RT-X Models101 citations · 2023