时间可解释可微决策树

中文摘要 AI 辅助

可解释性通过提供对智能体底层决策模型的透明性，为安全自主性提供了一种解决方案。在序列决策任务中，可微决策树（DDTs）是实现这种可解释性的一种方法，它在保持自动可微策略的同时，为人类提供基于离散树的可视化。然而，当前DDTs的实现并不适用于序列决策领域，因为树的单时间步行为与人类的多时间步规划之间存在固有的不匹配。因此，我们的工作将时间引入为可解释性的新维度，称为时间可解释性，并展示了通过动作分块（action chunking）实现的时间抽象如何提升这种可解释性。我们首先引入了两种结合动作分块的新型策略梯度算法来实现这一目标。此外，为了保持参数高效的树，我们开发了一种信息论树重构算法，该算法在训练过程中修改树。在四个模拟环境中，我们发现从蒸馏的动作分块策略热启动动作分块DDTs是获得时间可解释树的最有效方法：在四个领域中的三个领域，它们匹配了神经网络策略，同时使用的参数减少了高达80%。我们的代码可在以下https URL获取。

英文摘要

Interpretability offers a solution to safe autonomy by providing transparency into an agent's underlying decision-making model. Within sequential-decision making tasks, differentiable decision trees (DDTs) are one approach to such interpretability, maintaining automatic-differentiable policies while providing humans with a discrete tree-based visualization. Nonetheless, current implementations of DDTs are not well-suited for sequential-decision making domains, as there exists an inherent mismatch between a tree's single-timestep behavior and a human's multi-timestep planning. Our work thus introduces time as a new dimension of interpretability, coined as temporal interpretability, and demonstrates how temporal abstractions via action chunking improve it. We achieve this by first introducing two novel policy gradient algorithms that incorporate action chunking. Additionally, to maintain parameter-efficient trees, we develop an information-theoretic tree restructuring algorithm that modifies the tree during training. Across four simulation environments, we find that warm-starting action chunked DDTs from a distilled action chunked policy is the most effective way to obtain temporally interpretable trees: they match neural network policies in three of the four domains while using up to 80$\%$ fewer parameters. Our code is available at https://github.com/ei5uke/temp-interp.

发表机构

• Purdue University（普渡大学）
• University of Southern California（南加州大学）

机构由 AI 辅助整理，请以论文原文为准。