Branches：使用 AO* 高效寻找最优稀疏决策树
Branches: Efficiently Seeking Optimal Sparse Decision Trees with AO*
AI总结：
本文提出基于 AND/OR 图与 AO* 搜索的 Branches 算法，用于高效求解最优稀疏决策树，支持非二元特征，并给出最优性、复杂度及实验优势。
摘要：
决策树（Decision Tree，DT）学习是可解释机器学习中的一个基础问题，但它也带来了严峻的优化挑战。近期出现了一些实用算法，主要利用动态规划和分支定界。然而，这些方法大多依赖深度优先搜索策略，在搜索高深度 DT 时效率较低，并且需要定义最大深度超参数。其他方法也采用最佳优先搜索来规避这些问题。该策略的缺点是内存消耗较高，因此必须以完全高效的方式设计它，充分利用问题的结构。我们将该问题表述在 AND/OR 图搜索框架内，并用一种名为 Branches 的新型 AO* 类算法求解。我们证明了 Branches 的最优性和复杂度保证，并从理论上以及一系列实验中表明，它比现有最先进方法更高效。此外，与其他方法不同，Branches 支持非二元特征；我们表明，这一特性还能进一步带来更大的计算效率收益。
英文摘要：
Decision Tree (DT) Learning is a fundamental problem in Interpretable Machine Learning, yet it poses a formidable optimisation challenge. Practical algorithms have recently emerged, primarily leveraging Dynamic Programming and Branch & Bound. However, most of these approaches rely on a Depth-First-Search strategy, which is inefficient when searching for DTs at high depths and requires the definition of a maximum depth hyperparameter. Best-First-Search was also employed by other methods to circumvent these issues. The downside of this strategy is its higher memory consumption, as such, it has to be designed in a fully efficient manner that takes full advantage of the problem's structure. We formulate the problem within an AND/OR graph search framework and we solve it with a novel AO*-type algorithm called Branches. We prove both optimality and complexity guarantees for Branches and we show that it is more efficient than the state of the art theoretically and on a variety of experiments. Furthermore, Branches supports non-binary features unlike the other methods, we show that this property can further induce larger gains in computational efficiency.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
