paper-with-me

홈 › Papers

Goal Misgeneralization in Deep Reinforcement Learning

2021-05-28 · Lauro Langosco, Jack Koch, Lee Sharkey, Jacob Pfau, Laurent Orseau, David Krueger

We study goal misgeneralization, a type of out-of-distribution generalization failure in reinforcement learning (RL). Goal misgeneralization failures occur when an RL agent retains its capabilities out-of-distribution yet pursues the wrong goal. For instance, an agent might continue to competently avoid obstacles, but navigate to the wrong place. In contrast, previous works have typically focused on capability generalization failures, where an agent fails to do anything sensible at test time. We formalize this distinction between capability and goal generalization, provide the first empirical demonstrations of goal misgeneralization, and present a partial characterization of its causes.

📄 PDF Abstract BibTeX arXiv:2105.14111

Code (4)

JacobPfau/procgenAISC 공식 구현 pytorch
jbkjr/train-procgen-pytorch 공식 구현 pytorch
KarolisRam/colour-shape-goal-misgeneralization pytorch
UlisseMini/procgen-tools pytorch

Tasks

Deep Reinforcement LearningNavigateOut-of-Distribution Generalizationreinforcement-learningReinforcement LearningReinforcement Learning (RL)

Similar Papers 제목 키워드 기반

Mitigating Goal Misgeneralization via Minimax Regret

2025-07-03 · Karim Abdel Sadek, Matthew Farrugia-Roberts, Usman Anwar, Hannah Erlebach 외 arxiv

Safe generalization in reinforcement learning requires not only that a learned policy acts capably in new situations, but also that it uses its capabilities towards the pursuit of the designer's intended goal. The latter…

Reinforcement Learning

Reinforcement Learning from LLM Feedback to Counteract Goal Misgeneralization

2024-01-14 · Houda Nait El Barj, Theophile Sautory

We introduce a method to address goal misgeneralization in reinforcement learning (RL), leveraging Large Language Model (LLM) feedback during training. Goal misgeneralization, a type of robustness failure in RL occurs wh…

Language ModelingLanguage ModellingLarge Language Modelreinforcement-learning+2

Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals

2022-10-04 · Rohin Shah, Vikrant Varma, Ramana Kumar, Mary Phuong 외

The field of AI alignment is concerned with AI systems that pursue unintended goals. One commonly studied mechanism by which an unintended goal might arise is specification gaming, in which the designer-provided specific…

Getting By Goal Misgeneralization With a Little Help From a Mentor

2024-10-28 · Tu Trinh, Mohamad H. Danesh, Nguyen X. Khanh, Benjamin Plaut

While reinforcement learning (RL) agents often perform well during training, they can struggle with distribution shift in real-world deployments. One particularly severe risk of distribution shift is goal misgeneralizati…

Reinforcement Learning (RL)

Colour versus Shape Goal Misgeneralization in Reinforcement Learning: A Case Study

2023-12-05 · Karolis Ramanauskas, Özgür Şimşek

We explore colour versus shape goal misgeneralization originally demonstrated by Di Langosco et al. (2022) in the Procgen Maze environment, where, given an ambiguous choice, the agents seem to prefer generalization based…

reinforcement-learning