Marc Toussaint

About

Marc Toussaint is a leading robotics and artificial intelligence researcher whose work spans robot motion planning, task and motion planning (TAMP), and embodied AI. His research has fundamentally shaped how robots reason about and interact with the physical world, blending probabilistic inference, optimization, and symbolic reasoning into cohesive frameworks for intelligent robot control. Among his most influential contributions is his 2009 work reformulating stochastic optimal control as approximate inference for trajectory optimization, garnering 329 citations and reshaping how the field approached motion planning under uncertainty. His Logic-Geometric Programming framework (2015) offered a powerful optimization-based approach to combined task and motion planning, enabling robots to solve complex sequential manipulation problems. More recently, Toussaint contributed to the landmark PaLM-E project (2023, 350 citations), advancing embodied multimodal language models that ground large-scale AI reasoning directly in real-world sensory data — a significant step toward general-purpose robotic intelligence. His work on Gaussian process implicit surfaces, uncertainty-aware grasping, and multi-robot construction planning further demonstrates a consistent commitment to robust, uncertainty-aware manipulation. With hundreds of citations across diverse subfields, Toussaint stands as a foundational figure bridging classical robotics and modern machine learning.

Research Focus

Key Achievements

Top Papers

1. 1
   PaLM-E: An Embodied Multimodal Language Model

   350 citations · 2023
2. 2
   Robot trajectory optimization using approximate inference

   329 citations · 2009
3. 3
   Logic-geometric programming: an optimization-based approach to combined task and motion planning

   186 citations · 2015
4. 4
   Gaussian process implicit surfaces for shape estimation and grasping

   162 citations · 2011
5. 5
   Long-Horizon Multi-Robot Rearrangement Planning for Construction Assembly

   86 citations · 2022
6. 6
   Fast motion planning from experience: trajectory prediction for speeding up movement generation

   79 citations · 2013
7. 7
   Learning model-free robot control by a Monte Carlo EM algorithm

   72 citations · 2009
8. 8
   Uncertainty aware grasping and tactile exploration

   66 citations · 2013
9. 9
   Multi-bound tree search for logic-geometric programming in cooperative manipulation domains

   65 citations · 2017
10. 10
    Understanding the geometry of workspace obstacles in Motion Optimization

    58 citations · 2015

Key Collaborators

Contact & Links