让模型通过探索对称性学习物理规律，提升对未知物理的预测能力
DreamSAC: Learning Hamiltonian World Models via Symmetry Exploration
- 用哈密顿量激励的探索策略主动发现物理规律
- 在3D物理模拟中实现显著优于基线的外推性能
- 适合需要强物理泛化能力的研究者
学习型世界模型擅长插值泛化，但在面对新物理属性时外推能力差。这是因为它们学习的是统计相关性，而非环境背后的生成规则，如物理不变性和守恒律。我们提出，学习这些不变性是实现稳健外推的关键。为此，首先引入无监督探索策略「对称性探索」，代理通过基于哈密顿量的内在好奇心奖励，主动探测并挑战其对守恒律的理解，从而收集具有物理意义的数据。其次，设计基于哈密顿量的世界模型，利用新颖的自监督对比目标，从依赖视角的原始像素观测中识别出不变的物理状态。所提出的框架DreamSAC，在3D物理模拟任务中训练于此类主动采集数据，显著优于现有最先进基线，在需要外推的任务上表现突出。
原文摘要 · Abstract (English)
Learned world models excel at interpolative generalization but fail at extrapolative generalization to novel physical properties. This limitation arises because they learn statistical correlations rather than the environment's underlying generative rules, such as physical invariances and conservation laws. We argue that learning these invariances is key to robust extrapolation. To achieve this, we first introduce \textbf{Symmetry Exploration}, an unsupervised exploration strategy where an agent is intrinsically motivated by a Hamiltonian-based curiosity bonus to actively probe and challenge its understanding of conservation laws, thereby collecting physically informative data. Second, we design a Hamiltonian-based world model that learns from the collected data, using a novel self-supervised contrastive objective to identify the invariant physical state from raw, view-dependent pixel observations. Our framework, \textbf{DreamSAC}, trained on this actively curated data, significantly outperforms state-of-the-art baselines in 3D physics simulations on tasks requiring extrapolation.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。