PulseAugur
EN
LIVE 14:01:17

LLM-Evolved Pattern Generators Achieve Admissible Heuristics for Planning

By PulseAugur Editorial · [2 sources] ·

Researchers have developed a novel method for learning domain-dependent heuristics that ensure admissibility for optimal classical planning. This approach utilizes an LLM-driven evolutionary program-synthesis framework to generate programs that create pattern collections for planning tasks. These patterns are then combined admissibly, resulting in heuristics that offer comparable coverage to existing methods but with significantly faster state evaluation and negligible overhead. AI

IMPACT Introduces a novel approach to learning admissible heuristics for optimal planning, potentially improving efficiency and interpretability in AI planning systems.

RANK_REASON The cluster contains an academic paper detailing a new method for optimal classical planning.

Read on arXiv cs.AI →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

LLM-Evolved Pattern Generators Achieve Admissible Heuristics for Planning

COVERAGE [2]

  1. arXiv cs.AI TIER_1 English(EN) · Windy Phung, Dominik Drexler, Arnaud Lequen, Jendrik Seipp ·

    LLM-Evolved Pattern Generators for Optimal Classical Planning

    arXiv:2606.02438v1 Announce Type: new Abstract: Learned heuristics have recently become a competitive alternative to traditional domain-independent heuristics for satisficing planning. Existing approaches, however, focus on improving search guidance rather than guaranteeing admis…

  2. arXiv cs.AI TIER_1 English(EN) · Jendrik Seipp ·

    LLM-Evolved Pattern Generators for Optimal Classical Planning

    Learned heuristics have recently become a competitive alternative to traditional domain-independent heuristics for satisficing planning. Existing approaches, however, focus on improving search guidance rather than guaranteeing admissibility, which makes them unsuitable for optima…

RELATED ENTITIES

RELATED TOPICS