paper-with-me

Papers

Policy-Value Alignment and Robustness in Search-based Multi-Agent Learning

2023-01-27 · Niko A. Grupen, Michael Hanlon, Alexis Hao, Daniel D. Lee, Bart Selman

Large-scale AI systems that combine search and learning have reached super-human levels of performance in game-playing, but have also been shown to fail in surprising ways. The brittleness of such models limits their efficacy and trustworthiness in real-world deployments. In this work, we systematically study one such algorithm, AlphaZero, and identify two phenomena related to the nature of exploration. First, we find evidence of policy-value misalignment -- for many states, AlphaZero's policy and value predictions contradict each other, revealing a tension between accurate move-selection and value estimation in AlphaZero's objective. Further, we find inconsistency within AlphaZero's value function, which causes it to generalize poorly, despite its policy playing an optimal strategy. From these insights we derive VISA-VIS: a novel method that improves policy-value alignment and value robustness in AlphaZero. Experimentally, we show that our method reduces policy-value misalignment by up to 76%, reduces value generalization error by up to 50%, and reduces average value error by up to 55%.

📄 PDF Abstract BibTeX arXiv:2301.11857

Code (0)

등록된 구현이 없습니다.

Methods 이 논문이 사용한 방법론

fail 설명 없음
AlphaZero AlphaZero is a reinforcement learning agent for playing board games such as Go, chess, and shogi.

Similar Papers 제목 키워드 기반

Scaling World-Model Reinforcement Learning Through Diffusion Policy Optimization

2026-05-25 · Xiaoyuan Cheng, Wenxuan Yuan, Zhancun Mu, Yuanzhao Zhang 외 arxiv

Model-based reinforcement learning (RL) can be effectively supported at scale through the use of world models. However, in practice, scaling such approaches remains fundamentally limited. A commonly recognized challenge …

Reinforcement Learning

DVMap: Fine-Grained Pluralistic Value Alignment via High-Consensus Demographic-Value Mapping

2026-05-14 · Pengyun Zhu, Yuqi Ren, Zhen Wang, Lei Yang 외 arxiv

Current Large Language Models (LLMs) typically rely on coarse-grained national labels for pluralistic value alignment. However, such macro-level supervision often obscures intra-country value heterogeneity, yielding a lo…

Don't Forget Your Reward Values: Language Model Alignment via Value-based Calibration

2024-02-25 · Xin Mao, Feng-Lin Li, Huimin Xu, Wei zhang 외

While Reinforcement Learning from Human Feedback (RLHF) significantly enhances the generation quality of Large Language Models (LLMs), recent studies have raised concerns regarding the complexity and instability associat…

Language ModelingLanguage Modelling

Unifying Value Alignment and Assignment in Cross-Domain Offline Reinforcement Learning with Heterogeneous Datasets

2026-05-24 · Zhongjian Qiao, Jiafei Lyu, Chenjia Bai, Peisong Wang 외 arxiv

Cross-domain offline reinforcement learning (RL) aims to learn a policy in the target domain with a limited target domain dataset and a source domain dataset that exhibits a dynamics shift. Training directly on the origi…

Representation LearningReinforcement LearningOffline RL

D2VBench: Benchmarking Large Language Models with Value Dilemmas in Daily Scenarios

2026-07-22 · Siyi Hao, Yidi Cao, Linhao Yu, Yuqi Ren 외 arxiv

With the wide application of large language models (LLMs) in real-world scenarios, the value implication of their outputs is crucial. However, existing evaluation benchmarks suffer from insufficient coverage of value dil…