AutoSci: AI System for Full Scientific Research Lifecycle
AutoSci is a memory-centric agentic system designed to automate the entire scientific research lifecycle. It addresses the limitations of existing LLM-based scientific agents by supporting persistent structured memory across projects and self-improving research procedures. The system is organized into four modules, including SciMem which separates Long-Term Knowledge Memory from Active Research Memory. AutoSci aims to unify the research process from literature review to experiment management and manuscript preparation.
Key facts
- AutoSci is a memory-centric agentic system for the full scientific research lifecycle.
- It is based on arXiv paper 2605.31468.
- The system uses LLM-based scientific agents.
- It supports structured persistent memory across projects.
- AutoSci improves its own research procedures over time.
- It is organized around four modules.
- SciMem provides schema-governed research memory.
- SciMem separates Long-Term Knowledge Memory from Active Research Memory.
Entities
—