用解码器分析神经网络如何自动发现旅行商问题的几何规律
Mechanistic Interpretability for Neural TSP Solvers
- 通过稀疏自编码器解析模型内部激活，提取可解释特征
- 发现边界检测、聚类敏感和分割等几何特征，无需人工标注
- 为神经优化模型提供透明化路径，适合关注可解释性的研究者
神经网络在组合优化领域取得进展，基于Transformer的求解器能在毫秒级时间内获得接近最优的旅行商问题（TSP）解。然而，这些模型如黑箱运行，无法揭示其学习到的几何模式或构造路径时采用的启发式规则。本文首次将基于激活的可解释性方法——稀疏自编码器（SAEs）应用于基于Transformer的TSP求解器中，以解决这一问题。我们在100个节点的实例上使用强化学习训练了一个指针网络，随后在编码器残差流上拟合一个SAE，以发现一个过完备的可解释特征字典。分析表明，该求解器自然发展出与基本TSP概念相匹配的特征：在凸包节点上激活的边界检测器、对局部密集区域响应的聚类敏感特征，以及编码几何划分的分离特征。这些发现首次提供了神经TSP求解器在选择节点前内部计算的模型内解释，证明了几何结构可在无显式监督下自发出现，并为融合神经效率与算法可解释性的透明混合系统指明方向。交互式特征探索工具：https://reubennarad.github.io/TSP_interp
原文摘要 · Abstract (English)
Neural networks have advanced combinatorial optimization, with Transformer-based solvers achieving near-optimal solutions on the Traveling Salesman Problem (TSP) in milliseconds. However, these models operate as black boxes, providing no insight into the geometric patterns they learn or the heuristics they employ during tour construction. We address this opacity by applying sparse autoencoders (SAEs), a mechanistic interpretability technique, to a Transformer-based TSP solver, representing the first application of activation-based interpretability methods to operations research models. We train a pointer network with reinforcement learning on 100-node instances, then fit an SAE to the encoder's residual stream to discover an overcomplete dictionary of interpretable features. Our analysis reveals that the solver naturally develops features mirroring fundamental TSP concepts: boundary detectors that activate on convex-hull nodes, cluster-sensitive features responding to locally dense regions, and separator features encoding geometric partitions. These findings provide the first model-internal account of what neural TSP solvers compute before node selection, demonstrate that geometric structure emerges without explicit supervision, and suggest pathways toward transparent hybrid systems that combine neural efficiency with algorithmic interpretability. Interactive feature explorer: https://reubennarad.github.io/TSP_interp
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。