CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Siegel, Zachary S., Kapoor, Sayash, Nagdir, Nitya, Stroebl, Benedikt, Narayanan, Arvind
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929502330814464
author Siegel, Zachary S.
Kapoor, Sayash
Nagdir, Nitya
Stroebl, Benedikt
Narayanan, Arvind
author_facet Siegel, Zachary S.
Kapoor, Sayash
Nagdir, Nitya
Stroebl, Benedikt
Narayanan, Arvind
contents AI agents have the potential to aid users on a variety of consequential tasks, including conducting scientific research. To spur the development of useful agents, we need benchmarks that are challenging, but more crucially, directly correspond to real-world tasks of interest. This paper introduces such a benchmark, designed to measure the accuracy of AI agents in tackling a crucial yet surprisingly challenging aspect of scientific research: computational reproducibility. This task, fundamental to the scientific process, involves reproducing the results of a study using the provided code and data. We introduce CORE-Bench (Computational Reproducibility Agent Benchmark), a benchmark consisting of 270 tasks based on 90 scientific papers across three disciplines (computer science, social science, and medicine). Tasks in CORE-Bench consist of three difficulty levels and include both language-only and vision-language tasks. We provide an evaluation system to measure the accuracy of agents in a fast and parallelizable way, saving days of evaluation time for each run compared to a sequential implementation. We evaluated two baseline agents: the general-purpose AutoGPT and a task-specific agent called CORE-Agent. We tested both variants using two underlying language models: GPT-4o and GPT-4o-mini. The best agent achieved an accuracy of 21% on the hardest task, showing the vast scope for improvement in automating routine scientific tasks. Having agents that can reproduce existing work is a necessary step towards building agents that can conduct novel research and could verify and improve the performance of other research agents. We hope that CORE-Bench can improve the state of reproducibility and spur the development of future research agents.
format Preprint
id arxiv_https___arxiv_org_abs_2409_11363
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark
Siegel, Zachary S.
Kapoor, Sayash
Nagdir, Nitya
Stroebl, Benedikt
Narayanan, Arvind
Computation and Language
Artificial Intelligence
Multiagent Systems
AI agents have the potential to aid users on a variety of consequential tasks, including conducting scientific research. To spur the development of useful agents, we need benchmarks that are challenging, but more crucially, directly correspond to real-world tasks of interest. This paper introduces such a benchmark, designed to measure the accuracy of AI agents in tackling a crucial yet surprisingly challenging aspect of scientific research: computational reproducibility. This task, fundamental to the scientific process, involves reproducing the results of a study using the provided code and data. We introduce CORE-Bench (Computational Reproducibility Agent Benchmark), a benchmark consisting of 270 tasks based on 90 scientific papers across three disciplines (computer science, social science, and medicine). Tasks in CORE-Bench consist of three difficulty levels and include both language-only and vision-language tasks. We provide an evaluation system to measure the accuracy of agents in a fast and parallelizable way, saving days of evaluation time for each run compared to a sequential implementation. We evaluated two baseline agents: the general-purpose AutoGPT and a task-specific agent called CORE-Agent. We tested both variants using two underlying language models: GPT-4o and GPT-4o-mini. The best agent achieved an accuracy of 21% on the hardest task, showing the vast scope for improvement in automating routine scientific tasks. Having agents that can reproduce existing work is a necessary step towards building agents that can conduct novel research and could verify and improve the performance of other research agents. We hope that CORE-Bench can improve the state of reproducibility and spur the development of future research agents.
title CORE-Bench: Fostering the Credibility of Published Research Through a Computational Reproducibility Agent Benchmark
topic Computation and Language
Artificial Intelligence
Multiagent Systems
url https://arxiv.org/abs/2409.11363