AI-Researcher: Autonomous Scientific Innovation

Summary

HKUDS's end-to-end system covering literature review, hypothesis generation, algorithm implementation and manuscript preparation with minimal human intervention, together with Scientist-Bench, a benchmark of state-of-the-art papers across AI research domains with guided and open-ended tasks. The authors report high implementation success and papers approaching human-level quality. Code at github.com/HKUDS/AI-Researcher (about 5,800 stars, no licence declared, last pushed October 2025).

Contribution

Queued — not yet read.

Method

Queued — not yet read.

Relevance to RISE

A widely used end-to-end system not yet in the landscape. It is one of the baselines whose outputs [meng2026scientistone] audits, so its integrity profile is on record from an external source.

Critique / open questions

Queued — not yet assessed.

Key quotes

"To rigorously assess autonomous research capabilities, we develop Scientist-Bench, a comprehensive benchmark comprising state-of-the-art papers across diverse AI research domains, featuring both guided innovation and open-ended exploration tasks." (abstract)