Saved in:
Bibliographic Details
Main Authors: Huang, Xu, Chen, Junwu, Fei, Yuxing, Li, Zhuohan, Schwaller, Philippe, Ceder, Gerbrand
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2512.23880
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912853919793152
author Huang, Xu
Chen, Junwu
Fei, Yuxing
Li, Zhuohan
Schwaller, Philippe
Ceder, Gerbrand
author_facet Huang, Xu
Chen, Junwu
Fei, Yuxing
Li, Zhuohan
Schwaller, Philippe
Ceder, Gerbrand
contents Large language model (LLM) agents currently depend on predefined tools or early-stage tool generation, limiting their adaptability and scalability to complex scientific tasks. We introduce CASCADE, a self-evolving agentic framework representing an early instantiation of the transition from "LLM + tool use" to "LLM + skill acquisition". CASCADE enables agents to master complex external tools and codify knowledge through two meta-skills: continuous learning via web search, code extraction, and memory utilization; self-reflection via introspection, knowledge graph exploration, and others. We evaluate CASCADE on SciSkillBench, a benchmark of 116 materials science and chemistry research tasks. CASCADE achieves a 93.3% success rate using GPT-5, compared to 35.4% without evolution mechanisms. We further demonstrate real-world applications in computational analysis, autonomous laboratory experiments, and selective reproduction of published papers. Along with human-agent collaboration and memory consolidation, CASCADE accumulates executable skills that can be shared across agents and scientists, moving toward scalable AI-assisted scientific research.
format Preprint
id arxiv_https___arxiv_org_abs_2512_23880
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CASCADE: Cumulative Agentic Skill Creation through Autonomous Development and Evolution
Huang, Xu
Chen, Junwu
Fei, Yuxing
Li, Zhuohan
Schwaller, Philippe
Ceder, Gerbrand
Artificial Intelligence
Materials Science
Large language model (LLM) agents currently depend on predefined tools or early-stage tool generation, limiting their adaptability and scalability to complex scientific tasks. We introduce CASCADE, a self-evolving agentic framework representing an early instantiation of the transition from "LLM + tool use" to "LLM + skill acquisition". CASCADE enables agents to master complex external tools and codify knowledge through two meta-skills: continuous learning via web search, code extraction, and memory utilization; self-reflection via introspection, knowledge graph exploration, and others. We evaluate CASCADE on SciSkillBench, a benchmark of 116 materials science and chemistry research tasks. CASCADE achieves a 93.3% success rate using GPT-5, compared to 35.4% without evolution mechanisms. We further demonstrate real-world applications in computational analysis, autonomous laboratory experiments, and selective reproduction of published papers. Along with human-agent collaboration and memory consolidation, CASCADE accumulates executable skills that can be shared across agents and scientists, moving toward scalable AI-assisted scientific research.
title CASCADE: Cumulative Agentic Skill Creation through Autonomous Development and Evolution
topic Artificial Intelligence
Materials Science
url https://arxiv.org/abs/2512.23880