AutoSci: A Memory-Centric Agentic System for the Full Scientific Research Lifecycle

Researchers introduced AutoSci, a memory-centric AI agent system for end-to-end scientific research automation.

It combines long-term and project-level memory with lifecycle orchestration, multi-agent skill modules, and feedback-driven evolution.

The system is designed to support the full research workflow, from literature review and experiments to manuscript writing and rebuttal.

Its goal is to act as a unified AI research assistant that reuses knowledge across projects and improves over time.