Joshua B. Tenenbaum

About

Joshua B. Tenenbaum stands at the forefront of computational cognitive science and robot learning, weaving together physical simulation, neural representation, and embodied intelligence into a cohesive research vision. His work explores how machines can develop human-like intuitions about the physical world — understanding how objects move, feel, and interact across complex environments. Tenenbaum's contributions to differentiable physics simulation, most notably through frameworks like ChainQueen (223 citations), have empowered robots to solve inverse problems with remarkable efficiency by embedding physical simulators directly into gradient-based learning pipelines. His investigations into particle-based dynamics and neural physical representations (collectively exceeding 280 citations) demonstrate how flexible, hierarchical models can capture the rich behavior of rigid bodies, deformable objects, and fluids alike. Beyond simulation, Tenenbaum has advanced multimodal perception, integrating touch, vision, and shape priors to give robots richer environmental understanding. His work on Neural Descriptor Fields and ConceptGraphs (totaling over 316 citations) bridges 3D spatial reasoning with open-vocabulary semantics, enabling robots to perceive and plan across diverse manipulation tasks. Throughout, his research reflects a singular ambition: building machines that don't merely observe the world, but genuinely comprehend it — physically, semantically, and interactively.

Research Focus

Key Achievements

Top Papers

1. 1
   ChainQueen: A Real-Time Differentiable Physical Simulator for Soft Robotics

   223 citations · 2019
2. 2
   Neural Radiance Flow for 4D View Synthesis and Video Processing

   209 citations · 2021
3. 3
   ConceptGraphs: Open-Vocabulary 3D Scene Graphs for Perception and Planning

   178 citations · 2024
4. 4
   Neural Descriptor Fields: SE(3)-Equivariant Object Representations for Manipulation

   138 citations · 2022
5. 5
   See, feel, act: Hierarchical learning for complex manipulation skills with multisensory fusion

   124 citations · 2019
6. 6
   3D Shape Perception from Monocular Vision, Touch, and Shape Priors

   118 citations · 2018
7. 7
   Learning Particle Dynamics for Manipulating Rigid Bodies, Deformable Objects, and Fluids

   107 citations · 2018
8. 8
   Augmenting Physical Simulators with Stochastic Neural Networks: Case Study of Planar Pushing and Bouncing

   95 citations · 2018
9. 9
   Flexible Neural Representation for Physics Prediction

   91 citations · 2018
10. 10
    DensePhysNet: Learning Dense Physical Object Representations Via Multi-Step Dynamic Interactions

    87 citations · 2019

Key Collaborators

Contact & Links