What Matters in Hierarchical Search for Combinatorial Reasoning Problems?
This work addresses the challenge of efficiently solving NP-hard combinatorial reasoning problems for AI research, though it appears incremental as it focuses on refining existing subgoal methods rather than introducing a new paradigm.
The study investigated subgoal-planning methods for combinatorial reasoning problems, identifying key attributes like hard-to-learn value functions and complex action spaces that enable high-level search to outperform low-level planners, and proposed a consistent evaluation methodology to reevaluate state-of-the-art algorithms.
Efficiently tackling combinatorial reasoning problems, particularly the notorious NP-hard tasks, remains a significant challenge for AI research. Recent efforts have sought to enhance planning by incorporating hierarchical high-level search strategies, known as subgoal methods. While promising, their performance against traditional low-level planners is inconsistent, raising questions about their application contexts. In this study, we conduct an in-depth exploration of subgoal-planning methods for combinatorial reasoning. We identify the attributes pivotal for leveraging the advantages of high-level search: hard-to-learn value functions, complex action spaces, presence of dead ends in the environment, or using data collected from diverse experts. We propose a consistent evaluation methodology to achieve meaningful comparisons between methods and reevaluate the state-of-the-art algorithms.