Motion priors / short-horizon control / open source research

Motion priors for System 1 control.

I build models and open systems that turn motion data into reusable motion priors for robots: a learned System 1 reflex layer for the next few seconds of control.

Open source research / ProtoMotions

ProtoMotions is a shared framework for humanoid simulation and control, giving the community a practical starting point to experiment, build, and move faster.

2.1kcommunity stars Visit the framework ↗
Chen Tessler with family outdoors Family
field test
Chen TesslerResearch / systems / motion

01 Selected work

Motion render / 01MaskedMimic
Project page ↗

Physics-based robotics / zero-shot guidance

MaskedMimic

A zero-shot, guidable physics-based controller that turns partial intent into whole-body behavior across terrains, tasks, and environments.

Zero-shot guidanceMotion inpainting
Featured work
Research path

Research thread / reusable priors

One motion vocabulary, then more bodies and tasks.

  1. AnchorMaskedMimic

    Partial evidence becomes a reusable physics-based motion prior.

  2. ExtensionMaskedManipulator

    The same idea reaches into expressive whole-body manipulation.

  3. Next directionNew control modes, robots, scene awareness

    The same motion prior can extend to new forms of control and interaction.

Motion render / 02GPC
Project page ↗

Physics-based robotics / reusable motion prior

Generative Pretrained Controllers

A generative controller trained on large-scale motion data, then adapted to downstream tasks in physics simulation and robotics without rebuilding its movement vocabulary from scratch.

Reusable priorsTransferable skills
Featured work
Motion render / 03Kimodo
Project page ↗

Kinematic model / synthetic data

Kimodo

A controllable kinematic motion model for generating expressive human movement and the synthetic data needed to build with it.

Kinematic modelSynthetic data
Featured work

02 Publications

Motion generation, control, learning, and the systems around them.

26 research outputs / 2017-2026

01 / Archive group

Motion

Reusable priors, whole-body behavior, and the motion they generate.

11 papers

GPC: Large-Scale Generative Pretraining for Transferable Motor Control

Yi Shi, Yifeng Jiang, Chen Tessler, Xue Bin Peng

Pretraining a generative controller so learned motor behavior can transfer across robots, tasks, and new goals.

2026
SIGGRAPHPaperProjectCode

Kimodo: Scaling Controllable Human Motion Generation

Davis Rempe, Mathis Petrovich, Ye Yuan, Haotian Zhang, Xue Bin Peng, Yifeng Jiang, Tingwu Wang, Umar Iqbal, David Minor, Michael de Ruyter, Jiefeng Li, Chen Tessler, Edy Lim, Eugene Jeong, Sam Wu, Ehsan Hassani, Michael Huang, Jin-Bey Yu, Chaeyeon Chung, Lina Song, Olivier Dionne, Jan Kautz, Simon Yuen, Sanja Fidler

A scalable kinematic motion model that brings control, text, and constraints together into one expressive movement space.

2026
ARXIV PREPRINTPaperProjectCode

HIL: Hybrid Imitation Learning for Dynamic Athletic Control

Jiashun Wang, Yifeng Jiang, Haotian Zhang, Chen Tessler, Davis Rempe, Jessica Hodgins, Xue Bin Peng

Combining motion tracking and adversarial imitation to compose athletic skills that stay adaptable in new environments.

2026
ACM TOGPaperProject

Task Tokens: A Flexible Approach to Adapting Behavior Foundation Models

Ron Vainshtein, Zohar Rimon, Shie Mannor, Chen Tessler

Learning a compact task-specific signal that adapts a behavior foundation model without erasing its broader motion vocabulary.

2026
ICLRPaper

MaskedManipulator: Versatile Whole-Body Manipulation

Chen Tessler, Yifeng Jiang, Erwin Coumans, Zhengyi Luo, Gal Chechik, Xue Bin Peng

A masked motion-prior approach for extending spatio-temporal goals into whole-body manipulation.

2025
SIGGRAPH ASIAPaper

Emergent Active Perception and Dexterity of Simulated Humanoids from Visual Reinforcement Learning

Zhengyi Luo, Chen Tessler, Toru Lin, Ye Yuan, Tairan He, Wenli Xiao, Yunrong Guo, Gal Chechik, Kris Kitani, Linxi Fan, Yuke Zhu

Learning whole-body humanoid control from visual cues so search, gaze, and dexterous interaction emerge together.

2025
ARXIV PREPRINTPaperProject

MaskedMimic: Unified Physics-Based Character Control Through Masked Motion Inpainting

Chen Tessler, Yunrong Guo, Ofir Nabati, Gal Chechik, Jason Peng

Masked motion inpainting turns partial motion evidence into a reusable prior for physics-based character control.

2024
SIGGRAPH ASIAPaperProject

PlaMo: Plan and Move in Rich 3D Physical Environments

Assaf Hallak, Gal Dalal, Chen Tessler, Kelly Guo, Shie Mannor, Gal Chechik, Yunrong Guo

Pairing scene-aware path planning with a robust physics-based controller for rich, changing environments.

2024
ARXIV PREPRINTPaper

HumanoidOlympics: Sports Environments for Physically Simulated Humanoids

Zhengyi Luo, Jiashun Wang, Kangni Liu, Haotian Zhang, Chen Tessler, Jingbo Wang, Ye Yuan, Jinkun Cao, Zihui Lin, Fengyi Wang, Jessica K. Hodgins, Kris M. Kitani

A broad sports benchmark for developing human-like and performant humanoid behaviors across simulation and robotics.

2024
ARXIV PREPRINTPaperProject

CALM: Conditional Adversarial Latent Models for Directable Virtual Characters

Chen Tessler, Yoni Kasten, Yunrong Guo, Shie Mannor, Gal Chechik, Xue Bin Peng

A latent motion model that makes generated character behavior directable without giving up its natural variation.

2023
SIGGRAPHPaperProject

Action Robust Reinforcement Learning and Applications in Continuous Control

Chen Tessler, Yonathan Efroni, Shie Mannor

Designing policies that are robust when the action that reaches the world is not exactly the action that was intended.

2019
ICMLPaper

02 / Archive group

Learning

Algorithms, representations, and transferable behavior.

11 papers

Gradient Boosting Reinforcement Learning

Benjamin Fuhrer, Chen Tessler, Gal Dalal

Bringing the structure, efficiency, and interpretability of gradient-boosted trees into online reinforcement learning.

2025
ICMLPaperProjectCode

Learning to Move Like Professional Counter-Strike Players

David Durst, Feng Xie, Vishnu Sarukkai, Brennan Shacklett, Iuri Frosio, Chen Tessler, Joohwan Kim, Carly Taylor, Gilbert Bernstein, Sanjiban Choudhury, Pat Hanrahan, Kayvon Fatahalian

Learning the rhythm and spatial decisions of expert players as a controllable model of movement in a live environment.

2024
SCAPaper

Never Worse, Mostly Better: Stable Policy Improvement in Deep Reinforcement Learning

Pranav Khanna, Guy Tennenholtz, Nadav Merlis, Shie Mannor, Chen Tessler

Policy improvement with a stability guarantee: make progress without losing the behavior that already works.

2023
AAMASPaperProject

Ensemble Bootstrapping for Q-Learning

Oren Peer, Chen Tessler, Nadav Merlis, Ron Meir

Reducing Q-learning bias with an ensemble update that sits between over- and under-estimation.

2021
ICMLPaper

Inverse Reinforcement Learning in Contextual MDPs

Stav Belogolovsky, Philip Korsunsky, Shie Mannor, Chen Tessler, Tom Zahavy

Recovering the goals behind behavior when the environment itself changes with context.

2021
MACHINE LEARNINGPaperProject

Reward Tweaking: Maximizing the Total Reward While Planning for Short Horizons

Chen Tessler, Shie Mannor

Shaping a surrogate reward so short-horizon planning still optimizes the total reward that matters.

2020
ARXIV PREPRINTPaper

Distributional Policy Optimization: An Alternative Approach for Continuous Control

Chen Tessler, Guy Tennenholtz, Shie Mannor

Optimizing the distribution of outcomes directly to give continuous-control policies a richer view of what can happen next.

2019
NEURIPSPaperProject

Reward Constrained Policy Optimization

Chen Tessler, Daniel J. Mankowitz, Shie Mannor

A policy-optimization method for maximizing reward while respecting safety and resource constraints.

2019
ICLRPaper

Language is Power: Representing States Using Natural Language in Reinforcement Learning

Erez Schwartz, Guy Tennenholtz, Chen Tessler, Shie Mannor

Representing state semantically in language to give reinforcement-learning agents a more robust handle on complex situations.

2019
ARXIV PREPRINTPaper

Action Assembly: Sparse Imitation Learning for Text Based Games with Combinatorial Action Spaces

Chen Tessler, Tom Zahavy, Deborah Cohen, Daniel J. Mankowitz, Shie Mannor

Combining compressed sensing and imitation learning to make enormous discrete action spaces tractable.

2019
RLDMPaper

A Deep Hierarchical Approach to Lifelong Learning in Minecraft

Chen Tessler, Shahar Givony, Tom Zahavy, Daniel J. Mankowitz, Shie Mannor

Reusable skills enable learning long-horizon tasks in an open-ended world, where the next task is rarely known in advance.

2017
AAAIPaperProject

03 / Archive group

Applied

When learned ideas meet hardware, infrastructure, and design.

4 papers

Improving Inverse Folding for Peptide Design with Diversity-Regularized Direct Preference Optimization

Ryan Park, Darren J. Hsu, C. Brian Roland, Maria Korsunova, Chen Tessler, Shie Mannor, Olivia Viessmann, Bruno Trentini

Using preference optimization to make inverse-folding models produce more diverse peptide sequences without sacrificing structure.

2024
ARXIV PREPRINTPaper

Implementing Reinforcement Learning Datacenter Congestion Control in NVIDIA NICs

Benjamin Fuhrer, Yuval Shpigelman, Chen Tessler, Shie Mannor, Gal Chechik, Eitan Zahavy, Gal Dalal

Deploying learned congestion control in real time on NVIDIA NIC hardware inside a production datacenter.

2023
CCGRIDPaper

Reinforcement Learning for Datacenter Congestion Control

Chen Tessler, Yuval Shpigelman, Gal Dalal, Amit Mendelbaum, Doron Kazakov, Benjamin Fuhrer, Gal Chechik, Shie Mannor

Applying reinforcement learning to keep datacenter traffic moving under changing demand and tight latency budgets.

2022
IAAIPaper

Towards Autonomous Grading In The Real World

Yakov Miron, Chana Ross, Yuval Goldfracht, Chen Tessler, Dotan Di Castro

Bridging simulation and a scaled physical prototype to teach a dozer how to flatten uneven terrain.

2022
IROSPaper