AlphaResearch: Accelerating New Algorithm Discovery with Language Models

TL;DR

AlphaResearch is an autonomous agent that leverages language models to discover new algorithms for open-ended problems, outperforming existing systems and surpassing human performance on some tasks.

Contribution

The paper introduces AlphaResearch, a novel autonomous research agent with a dual environment for algorithm discovery, and provides a new benchmark dataset for evaluation.

Findings

AlphaResearch outperforms other discovery systems on six open-ended problems.

It achieves the best-known performance on the 'packing circles' problem, surpassing humans.

The study offers insights into the benefits and challenges of autonomous research agents.

Abstract

LLMs have made significant progress in complex but easy-to-verify problems, yet they still struggle with discovering the unknown. In this paper, we present \textbf{AlphaResearch}, an autonomous research agent designed to discover new algorithms on open-ended problems by iteratively running the following steps: (1) propose new ideas (2) program to verify (3) optimize the research proposals. To synergize the feasibility and innovation of the discovery process, we construct a novel dual environment by combining the execution-based verifiable reward and reward from simulated real-world peer review environment in AlphaResearch. We construct \textbf{\dataset}, a set of questions that includes an eight open-ended algorithmic problems competition to benchmark AlphaResearch. Experimental results show that AlphaResearch achieves stronger discovery performance than other agentic discovery systems on six open-ended problems. Notably, the algorithm discovered by AlphaResearch on the \emph{``packing circles''} problem achieves the best-of-known performance, surpassing the results of human researchers and strong baselines from recent work (e.g., AlphaEvolve). Additionally, we conduct a comprehensive analysis of the benefits and remaining challenges of autonomous research agent, providing valuable insights for future research.