Teaching Language Models to Forecast Research Success Through Comparative Idea Evaluation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mule, Srujan P, Garikaparthi, Aniketh, Patwardhan, Manasi
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918515566444544
author Mule, Srujan P
Garikaparthi, Aniketh
Patwardhan, Manasi
author_facet Mule, Srujan P
Garikaparthi, Aniketh
Patwardhan, Manasi
contents As language models accelerate scientific research by automating hypothesis generation and implementation, a new bottleneck emerges: evaluating and filtering hundreds of AI-generated ideas without exhaustive experimentation. We ask whether LMs can learn to forecast the empirical success of research ideas before any experiments are run. We study comparative empirical forecasting: given a benchmark-specific research goal and two candidate ideas, predict which will achieve better benchmark performance. We construct a dataset of 11,488 idea pairs grounded in objective outcomes from PapersWithCode. While off-the-shelf 8B-parameter models struggle (30% acc.), SFT dramatically boosts performance to 77.1%, outperforming GPT-5 (61.1%). By framing evaluation as a reasoning task via Reinforcement Learning with Verifiable Rewards (RLVR), we train models to discover latent reasoning paths, achieving 71.35% acc. with interpretable justifications. Through additional ablations and out-of-distribution tests, we show robustness to surface-level heuristics and transfer to both a cross-domain time-split test set and an independently constructed test set. Our results demonstrate that compute-efficient small language models can serve as effective, objective verifiers, offering a scalable path for autonomous scientific discovery.
format Preprint
id arxiv_https___arxiv_org_abs_2605_21491
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Teaching Language Models to Forecast Research Success Through Comparative Idea Evaluation
Mule, Srujan P
Garikaparthi, Aniketh
Patwardhan, Manasi
Machine Learning
Artificial Intelligence
Computation and Language
As language models accelerate scientific research by automating hypothesis generation and implementation, a new bottleneck emerges: evaluating and filtering hundreds of AI-generated ideas without exhaustive experimentation. We ask whether LMs can learn to forecast the empirical success of research ideas before any experiments are run. We study comparative empirical forecasting: given a benchmark-specific research goal and two candidate ideas, predict which will achieve better benchmark performance. We construct a dataset of 11,488 idea pairs grounded in objective outcomes from PapersWithCode. While off-the-shelf 8B-parameter models struggle (30% acc.), SFT dramatically boosts performance to 77.1%, outperforming GPT-5 (61.1%). By framing evaluation as a reasoning task via Reinforcement Learning with Verifiable Rewards (RLVR), we train models to discover latent reasoning paths, achieving 71.35% acc. with interpretable justifications. Through additional ablations and out-of-distribution tests, we show robustness to surface-level heuristics and transfer to both a cross-domain time-split test set and an independently constructed test set. Our results demonstrate that compute-efficient small language models can serve as effective, objective verifiers, offering a scalable path for autonomous scientific discovery.
title Teaching Language Models to Forecast Research Success Through Comparative Idea Evaluation
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2605.21491