理解神经路由求解器中的决策机制

中文摘要 AI 辅助

神经组合优化（NCO）已取得显著的实证成功，然而驱动模型决策的内部机制在很大程度上仍未被探索。在本文中，我们研究了跨越两种编码器-解码器配置的三个具有代表性的自回归NCO模型：AM和POMO（重编码器、轻解码器），以及LEHD（轻编码器、重解码器）。通过行为分析、表示探针和因果干预，我们考察了这些模型在解码过程中如何构建解决方案以及如何使用内部表示。我们的结果表明，AM和POMO在解决方案构建过程中主要遵循一种持续的几何模式，而LEHD包含关于多个未来动作的线性可访问信息。因果实验进一步为未来节点表示在LEHD决策中的作用提供了证据。我们还观察到，LEHD强烈依赖当前节点表示来进行即时局部决策，而起始节点表示在后续路径中扮演更广泛的导航角色。跨实例对齐分析还表明，LEHD将当前节点表示映射到一个相对共享的潜在区域，这可能为评估后续决策提供稳定的参考。在旅行商问题和容量受限车辆路径问题中，这些结果揭示了这些架构上不同的求解器之间的不同决策模式，并为NCO求解器的更可解释分析提供了基础。代码和额外的可视化在此https URL中提供。

英文摘要

Neural Combinatorial Optimization (NCO) has achieved strong empirical success, yet the internal mechanisms driving model decisions remain largely unexplored. In this paper, we investigate three representative autoregressive NCO models spanning two encoder-decoder configurations: AM and POMO (heavy-encoder, light-decoder), and LEHD (light-encoder, heavy-decoder). Through behavioral analyses, representation probing, and causal interventions, we examine how these models construct solutions and use internal representations during decoding. Our results suggest that AM and POMO predominantly follow a persistent geometric pattern throughout solution construction, whereas LEHD contains linearly accessible information about multiple future actions. Causal experiments further provide evidence for the role of future-node representations in LEHD's decision-making. We also observe that LEHD relies strongly on the current-node representation for immediate local decisions, while the start-node representation plays a broader navigational role over the subsequent route. Cross-instance alignment analyses additionally indicate that LEHD maps current-node representations into a relatively shared latent region, which may provide a stable reference for evaluating subsequent decisions. Across the Traveling Salesman Problem and the Capacitated Vehicle Routing Problem, these results reveal distinct decision-making patterns across these architecturally distinct solvers and provide a foundation for more interpretable analyses of NCO solvers. Code and additional visualizations are provided in the https://github.com/NCO-Interpretability/NCO-Interpretability.

发表机构

• Sharif University of Technology（谢里夫科技大学）

机构由 AI 辅助整理，请以论文原文为准。