Glen Berseth

Assistant Professor@University of Montreal, University of Montreal·加拿大·OpenReview

研究方向

Robotics · Deep Learning · artificial inteligence · multi-agent systems · planning · Reinforcement learning · physics-based animation · machine learning · Planning · Computer animation · motion planning · derivative free optimization

Glen Berseth

Temporal Representations for Exploration: Learning Complex Exploratory Behavior without Extrinsic Rewards

ARM-FM: Automated Reward Machines via Foundation Models for Compositional Reinforcement Learning

A Comedy of Estimators: On KL Regularization in RL Training of LLMs

Self-Predictive Representations for Combinatorial Generalization in Behavioral Cloning

Recursive Self-Aggregation Unlocks Deep Thinking in Large Language Models

Align and Filter: Improving Performance in Asynchronous On-Policy RL

M$^{4}$olGen: Multi-Agent, Multi-Stage Molecular Generation under Precise Multi-Property Constraints

Generalization in Online Reinforcement Learning for Mobile Agents

Is Exploration or Optimization the Problem for Deep Reinforcement Learning?

SegDAC: Improving Visual Reinforcement Learning by Extracting Dynamic Objectc-Centric Representations from Pretrained Vision Models

Concept-Based Steering of LLMs for Conditional Molecular Generation

CURATE: Automatic Curriculum Learning for Reinforcement Learning Agents through Competence-Based Curriculum Policy Search

Stable Gradients for Stable Learning at Scale in Deep Reinforcement Learning

Enabling Realtime Reinforcement Learning at Scale with Staggered Asynchronous Inference

Mitigating Plasticity Loss in Continual Reinforcement Learning by Reducing Churn

Outsourced Diffusion Sampling: Efficient Posterior Inference in Latent Spaces of Generative Models

Towards Improving Exploration through Sibling Augmented GFlowNets

Non-Adversarial Inverse Reinforcement Learning via Successor Feature Matching

EDiSon: Efficient Design-and-Control Optimization with Reinforcement Learning and Adaptive Design Reuse