SkillGen: Learning Domain Skills for In-Context Sequential Decision Making

2025-11-18

AI Summary

Large language models (LLMs) rely on high-quality in-context learning (ICL) prompts for sequential decision-making, yet existing approaches struggle to simultaneously achieve critical information focus, step-level granularity, and annotation efficiency. Method: We propose a skill-driven ICL framework that constructs an action-centric domain skill graph and integrates temporal-difference credit assignment to automatically identify critical decision paths. Leveraging trajectory sampling, skill retrieval, and fine-grained prompt generation, our method produces context prompts with high information density and minimal annotation overhead. Contribution/Results: This work is the first to theoretically unify domain-level skill graphs with credit assignment, ensuring task identifiability and guiding optimal prompt design. Experiments on ALFWorld, BabyAI, and ScienceWorld demonstrate average task completion rate improvements of 5.9–16.5% over state-of-the-art methods, significantly enhancing LLMs’ sequential decision-making capabilities in complex environments.

Technology Category

• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Large language models (LLMs) are increasingly applied to sequential decision-making through in-context learning (ICL), yet their effectiveness is highly sensitive to prompt quality. Effective prompts should meet three principles: focus on decision-critical information, provide step-level granularity, and minimize reliance on expert annotations through label efficiency. However, existing ICL methods often fail to satisfy all three criteria simultaneously. Motivated by these challenges, we introduce SkillGen, a skill-based ICL framework for structured sequential reasoning. It constructs an action-centric, domain-level graph from sampled trajectories, identifies high-utility actions via temporal-difference credit assignment, and retrieves step-wise skills to generate fine-grained, context-aware prompts. We further present a theoretical analysis showing that focusing on high-utility segments supports task identifiability and informs more effective ICL prompt design. Experiments on ALFWorld, BabyAI, and ScienceWorld, using both open-source and proprietary LLMs, show that SkillGen achieves consistent gains, improving progress rate by 5.9%-16.5% on average across models.

Problem

• Improves in-context learning for sequential decision making with LLMs
• Addresses prompt sensitivity by generating fine-grained skill-based prompts
• Reduces reliance on expert annotations through label-efficient methods

Innovation

• Constructs action-centric domain graph from trajectories
• Identifies high-utility actions via temporal-difference credit
• Retrieves step-wise skills for context-aware prompts