Karol Hausman

About

Karol Hausman is a leading robotics researcher whose work sits at the intersection of machine learning, large language models, and embodied AI. Best known for pioneering efforts to bridge the gap between powerful language and vision models and real-world robotic control, Hausman has helped define a new paradigm in which internet-scale knowledge can be transferred directly into physical systems. His landmark contributions include "Do As I Can, Not As I Say" (516 citations), which introduced the concept of grounding language model reasoning in robotic affordances, and "Code as Policies" (561 citations), which demonstrated that LLMs can author executable robot control programs from natural language commands. His work on RT-1 and RT-2 (512 and 267 citations, respectively) established transformer-based architectures as viable foundations for scalable, generalizable robot learning, while PaLM-E (350 citations) extended multimodal reasoning into embodied settings. Earlier contributions include Meta-World (282 citations), a widely adopted benchmark for meta- and multi-task reinforcement learning, and foundational work on tactile sensing and skill embedding. With over 3,200 citations across his most influential papers alone, Hausman's research has become essential reading for anyone studying the future of intelligent, language-guided robotics.

Research Focus

Key Achievements

Top Papers

1. 1
   Code as Policies: Language Model Programs for Embodied Control

   561 citations · 2023
2. 2
   Do As I Can, Not As I Say: Grounding Language in Robotic Affordances

   516 citations · 2022
3. 3
   RT-1: Robotics Transformer for Real-World Control at Scale

   512 citations · 2023
4. 4
   PaLM-E: An Embodied Multimodal Language Model

   350 citations · 2023
5. 5
   Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning

   282 citations · 2019
6. 6
   RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control

   267 citations · 2023
7. 7
   Force estimation and slip detection/classification for grip control using a biomimetic tactile sensor

   215 citations · 2015
8. 8
   Inner Monologue: Embodied Reasoning through Planning with Language Models

   206 citations · 2022
9. 9
   Learning an Embedding Space for Transferable Robot Skills

   190 citations · 2018
10. 10
    Foundation models in robotics: Applications, challenges, and the future

    163 citations · 2024

Key Collaborators

Contact & Links