HyperSkill: Self-Evolving LLM Agents via Hypergraph-Structured Skill Memory
2026-08-17 • Computation and Language
Computation and Language
AI summaryⓘ
The authors address how smart AI agents remember and reuse skills from previous tasks to solve harder problems. They point out that current methods store memories separately and retrieve them without considering connections between subtasks and skills. Their solution, called HyperSkill, uses a special memory map called a hypergraph linking steps and skills from entire task sequences. This helps the AI find useful skills better and keep its memory organized by removing less useful parts. Tests show HyperSkill works better than existing memory systems on several benchmarks involving popular AI models.
Large Language Models (LLMs)Experiential memoryHypergraphSubtasksReusable skillsMemory retrievalMemory maintenanceTrajectoryEmbedding similarityQA benchmarks
Authors
Ruiyao Xu, Tiankai Yang, Wei-Chieh Huang
Abstract
As agentic tasks grow in complexity, LLM agents increasingly rely on experiential memory to reuse procedural knowledge across tasks. Effective memory design must jointly address what to store, how memory is structured and retrieved, and how memory evolves. Existing systems tackle each only partially: they store trajectories, insights, or workflows as isolated entries, discarding compositional relationships among subtasks and reusable skills; retrieve by flat embedding similarity that ignores relational signals; and maintain memory without leveraging its relational structure. We propose HyperSkill, a hypergraph-based memory framework that jointly improves all three. HyperSkill represents memory as a hypergraph with two node types, subtask steps and reusable skills, where each hyperedge links the subtasks and skills from a single trajectory. Dual-path retrieval queries both subtask and trajectory levels, ranking skills by co-occurrence across retrieved trajectories. Periodic structure-informed maintenance prunes low-utility nodes and merges redundant skills via quality-weighted propagation. Across xBench, GAIA, and WebWalkerQA with GPT-4o and Qwen3-30B-A3B, HyperSkill outperforms ten memory baselines, yielding gains of up to +11.51 on GAIA and +11.18 on WebWalkerQA.