Chris Paxton

About

Chris Paxton is a robotics researcher whose work sits at the intersection of human-robot interaction, task planning, and the application of machine learning to real-world robotic systems. His research has consistently focused on making robots more accessible and useful to non-expert users — a theme evident from his early work on CoSTAR (2017, 170 citations), a groundbreaking system enabling end users to instruct collaborative robots through behavior trees and vision, and a complementary 2015 framework targeting small manufacturing environments (99 citations). Paxton has made significant contributions to human-robot handovers, developing vision-based systems capable of handling arbitrary objects with diverse shapes and deformability (83 citations), as well as grasp classification methods that underpin reactive collaboration (53 citations). More recently, his research has embraced large language models and foundation models, producing influential work on object rearrangement through commonsense reasoning (124 citations), semantic scene representations via CLIP-Fields (68 citations), and natural language plan correction (67 citations). His 2024 review of foundation models in real-world robotics (60 citations) reflects his broader role in synthesizing emerging trends for the field. With over 800 cumulative citations, Paxton stands as a leading voice in making intelligent, collaborative robots a practical reality.

Research Focus

Key Achievements

Top Papers

1. 1
   CoSTAR: Instructing collaborative robots with behavior trees and vision

   170 citations · 2017
2. 2
   Task and Motion Planning with Large Language Models for Object Rearrangement

   124 citations · 2023
3. 3
   A framework for end-user instruction of a robot assistant for manufacturing

   99 citations · 2015
4. 4
   Reactive Human-to-Robot Handovers of Arbitrary Objects

   83 citations · 2021
5. 5
   CLIP-Fields: Weakly Supervised Semantic Fields for Robotic Memory

   68 citations · 2023
6. 6
   Correcting Robot Plans with Natural Language Feedback

   67 citations · 2022
7. 7
   Real-world robot applications of foundation models: a review

   60 citations · 2024
8. 8
   StructFormer: Learning Spatial Structure for Language-Guided Semantic Rearrangement of Novel Objects

   55 citations · 2022
9. 9
   Human Grasp Classification for Reactive Human-to-Robot Handovers

   53 citations · 2020
10. 10
    An incremental approach to learning generalizable robot tasks from human demonstration

    49 citations · 2015

Key Collaborators

Contact & Links