Matthew Walter

Associate Professor@Toyota Technological Institute at Chicago·美国·OpenReview

研究方向

deep learning · reinforcement learning · Natural language grounding · Natural language generation · robot manipulation · human-robot interaction · robot learning · robotics · computer vision · SLAM

6.3

MANGO: A Benchmark for Evaluating Mapping and Navigation Abilities of Large Language Models

COLM 2024Poster

通讯

6.3

MANGO: A Benchmark for Evaluating Mapping and Navigation Abilities of Large Language Models

ICLR 2024Rejected

6.0

Subwords as Skills: Tokenization for Sparse-Reward Reinforcement Learning

NeurIPS 2024Poster

通讯

5.0

Subwords as Skills: Tokenization for Sparse-Reward Reinforcement Learning

合作者 (20)

Matthew Walter

Approaching Deep Learning through the Spectral Dynamics of Weights

PAC Learning with Improvements

Learning Actionable Counterfactual Explanations in Large State Spaces

Blending Imitation and Reinforcement Learning for Robust Policy Improvement

MANGO: A Benchmark for Evaluating Mapping and Navigation Abilities of Large Language Models

MANGO: A Benchmark for Evaluating Mapping and Navigation Abilities of Large Language Models

Subwords as Skills: Tokenization for Sparse-Reward Reinforcement Learning

Subwords as Skills: Tokenization for Sparse-Reward Reinforcement Learning