Chen Tessler
YOU?
Author Swipe
View article: MaskedManipulator: Versatile Whole-Body Manipulation
MaskedManipulator: Versatile Whole-Body Manipulation Open
We tackle the challenges of synthesizing versatile, physically simulated human motions for full-body object manipulation. Unlike prior methods that are focused on detailed motion tracking, trajectory following, or teleoperation, our framew…
View article: HIL: Hybrid Imitation Learning of Diverse Parkour Skills from Videos
HIL: Hybrid Imitation Learning of Diverse Parkour Skills from Videos Open
Recent data-driven methods leveraging deep reinforcement learning have been an effective paradigm for developing controllers that enable physically simulated characters to produce natural human-like behaviors. However, these data-driven me…
View article: Emergent Active Perception and Dexterity of Simulated Humanoids from Visual Reinforcement Learning
Emergent Active Perception and Dexterity of Simulated Humanoids from Visual Reinforcement Learning Open
Human behavior is fundamentally shaped by visual perception -- our ability to interact with the world depends on actively gathering relevant information and adapting our movements accordingly. Behaviors like searching for objects, reaching…
View article: Task Tokens: A Flexible Approach to Adapting Behavior Foundation Models
Task Tokens: A Flexible Approach to Adapting Behavior Foundation Models Open
Recent advancements in imitation learning have led to transformer-based behavior foundation models (BFMs) that enable multi-modal, human-like control for humanoid agents. While excelling at zero-shot generation of robust behaviors, BFMs of…
View article: MaskedMimic: Unified Physics-Based Character Control Through Masked Motion Inpainting
MaskedMimic: Unified Physics-Based Character Control Through Masked Motion Inpainting Open
Crafting a single, versatile physics-based controller that can breathe life into interactive characters across a wide spectrum of scenarios represents an exciting frontier in character animation. An ideal controller should support diverse …
View article: Improving Inverse Folding for Peptide Design with Diversity-regularized Direct Preference Optimization
Improving Inverse Folding for Peptide Design with Diversity-regularized Direct Preference Optimization Open
Inverse folding models play an important role in structure-based design by predicting amino acid sequences that fold into desired reference structures. Models like ProteinMPNN, a message-passing encoder-decoder model, are trained to reliab…
View article: MaskedMimic: Unified Physics-Based Character Control Through Masked Motion Inpainting
MaskedMimic: Unified Physics-Based Character Control Through Masked Motion Inpainting Open
Crafting a single, versatile physics-based controller that can breathe life into interactive characters across a wide spectrum of scenarios represents an exciting frontier in character animation. An ideal controller should support diverse …
View article: Gradient Boosting Reinforcement Learning
Gradient Boosting Reinforcement Learning Open
We present Gradient Boosting Reinforcement Learning (GBRL), a framework that adapts the strengths of gradient boosting trees (GBT) to reinforcement learning (RL) tasks. While neural networks (NNs) have become the de facto choice for RL, th…
View article: SMPLOlympics: Sports Environments for Physically Simulated Humanoids
SMPLOlympics: Sports Environments for Physically Simulated Humanoids Open
We present SMPLOlympics, a collection of physically simulated environments that allow humanoids to compete in a variety of Olympic sports. Sports simulation offers a rich and standardized testing ground for evaluating and improving the cap…
View article: PlaMo: Plan and Move in Rich 3D Physical Environments
PlaMo: Plan and Move in Rich 3D Physical Environments Open
Controlling humanoids in complex physically simulated worlds is a long-standing challenge with numerous applications in gaming, simulation, and visual content creation. In our setup, given a rich and complex 3D scene, the user provides a l…
View article: CALM: Conditional Adversarial Latent Models for Directable Virtual Characters
CALM: Conditional Adversarial Latent Models for Directable Virtual Characters Open
In this work, we present Conditional Adversarial Latent Models (CALM), an approach for generating diverse and directable behaviors for user-controlled interactive virtual characters. Using imitation learning, CALM learns a representation o…
View article: Implementing Reinforcement Learning Datacenter Congestion Control in NVIDIA NICs
Implementing Reinforcement Learning Datacenter Congestion Control in NVIDIA NICs Open
As communication protocols evolve, datacenter network utilization increases. As a result, congestion is more frequent, causing higher latency and packet loss. Combined with the increasing complexity of workloads, manual design of congestio…
View article: Implementing Reinforcement Learning Datacenter Congestion Control in NVIDIA NICs
Implementing Reinforcement Learning Datacenter Congestion Control in NVIDIA NICs Open
As communication protocols evolve, datacenter network utilization increases. As a result, congestion is more frequent, causing higher latency and packet loss. Combined with the increasing complexity of workloads, manual design of congestio…
View article: Reinforcement Learning for Datacenter Congestion Control
Reinforcement Learning for Datacenter Congestion Control Open
We approach the task of network congestion control in datacenters using Reinforcement Learning (RL). Successful congestion control algorithms can dramatically improve latency and overall network throughput. Until today, no such learning-ba…
View article: Towards Autonomous Grading In The Real World
Towards Autonomous Grading In The Real World Open
In this work, we aim to tackle the problem of autonomous grading, where a dozer is required to flatten an uneven area. In addition, we explore methods for bridging the gap between a simulated environment and real scenarios. We design both …
View article: Ensemble Bootstrapping for Q-Learning
Ensemble Bootstrapping for Q-Learning Open
Q-learning (QL), a common reinforcement learning algorithm, suffers from over-estimation bias due to the maximization term in the optimal Bellman operator. This bias may lead to sub-optimal behavior. Double-Q-learning tackles this issue by…
View article: Reinforcement Learning for Datacenter Congestion Control
Reinforcement Learning for Datacenter Congestion Control Open
We approach the task of network congestion control in datacenters using Reinforcement Learning (RL). Successful congestion control algorithms can dramatically improve latency and overall network throughput. Until today, no such learning-ba…
View article: Reward Tweaking: Maximizing the Total Reward While Planning for Short Horizons
Reward Tweaking: Maximizing the Total Reward While Planning for Short Horizons Open
In reinforcement learning, the discount factor $γ$ controls the agent's effective planning horizon. Traditionally, this parameter was considered part of the MDP; however, as deep reinforcement learning algorithms tend to become unstable wh…
View article: Maximizing the Total Reward via Reward Tweaking.
Maximizing the Total Reward via Reward Tweaking. Open
In reinforcement learning, the discount factor $\gamma$ controls the agent's effective planning horizon. Traditionally, this parameter was considered part of the MDP; however, as deep reinforcement learning algorithms tend to become unstab…
View article: Never Worse, Mostly Better: Stable Policy Improvement in Deep Reinforcement Learning
Never Worse, Mostly Better: Stable Policy Improvement in Deep Reinforcement Learning Open
In recent years, there has been significant progress in applying deep reinforcement learning (RL) for solving challenging problems across a wide variety of domains. Nevertheless, convergence of various methods has been shown to suffer from…
View article: Language is Power: Representing States Using Natural Language in Reinforcement Learning
Language is Power: Representing States Using Natural Language in Reinforcement Learning Open
Recent advances in reinforcement learning have shown its potential to tackle complex real-life tasks. However, as the dimensionality of the task increases, reinforcement learning methods tend to struggle. To overcome this, we explore metho…
View article: Stabilizing Deep Reinforcement Learning with Conservative Updates
Stabilizing Deep Reinforcement Learning with Conservative Updates Open
In recent years, advances in deep learning have enabled the application of reinforcement learning algorithms in complex domains. However, they lack the theoretical guarantees which are present in the tabular setting and suffer from many st…
View article: Stabilizing Off-Policy Reinforcement Learning with Conservative Policy Gradients
Stabilizing Off-Policy Reinforcement Learning with Conservative Policy Gradients Open
In recent years, advances in deep learning have enabled the application of reinforcement learning algorithms in complex domains. However, they lack the theoretical guarantees which are present in the tabular setting and suffer from many st…
View article: Natural Language State Representation for Reinforcement Learning
Natural Language State Representation for Reinforcement Learning Open
Recent advances in Reinforcement Learning have highlighted the difficulties in learning within complex high dimensional domains. We argue that one of the main reasons that current approaches do not perform well, is that the information is …
View article: Distributional Policy Optimization: An Alternative Approach for Continuous Control
Distributional Policy Optimization: An Alternative Approach for Continuous Control Open
We identify a fundamental problem in policy gradient-based methods in continuous control. As policy gradient methods require the agent's underlying probability distribution, they limit policy representation to parametric distribution class…
View article: Action Assembly: Sparse Imitation Learning for Text Based Games with Combinatorial Action Spaces
Action Assembly: Sparse Imitation Learning for Text Based Games with Combinatorial Action Spaces Open
We propose a computationally efficient algorithm that combines compressed sensing with imitation learning to solve text-based games with combinatorial action spaces. Specifically, we introduce a new compressed sensing algorithm, named IK-O…
View article: Learning personalized treatments via IRL
Learning personalized treatments via IRL Open
We consider the task of Inverse Reinforcement Learning in Contextual Markov Decision Processes (MDPs). In this setting, contexts that define the reward and transition kernel, are sampled from a distribution. Although the reward is a functi…
View article: Action Robust Reinforcement Learning and Applications in Continuous Control
Action Robust Reinforcement Learning and Applications in Continuous Control Open
A policy is said to be robust if it maximizes the reward while considering a bad, or even adversarial, model. In this work we formalize two new criteria of robustness to action uncertainty. Specifically, we consider two scenarios in which …