Evaluation-driven Scaling for Scientific Discovery

• URL: http://arxiv.org/abs/2604.19341v1
• Date: Tue, 21 Apr 2026 11:24:09 GMT
• Title: Evaluation-driven Scaling for Scientific Discovery
• Abstract summary: We introduce Simple Test-time Evaluation-driven Scaling (SimpleTES), a framework that combines parallel exploration, feedback-driven refinement, and local selection.<n>We find substantial gains unlocked by scaling evaluation-driven discovery loops along the right dimensions.<n>Our results establish effective evaluation-driven loop scaling as a central axis for advancing scientific discovery.
• Score: 77.20820317940581
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Language models are increasingly used in scientific discovery to generate hypotheses, propose candidate solutions, implement systems, and iteratively refine them. At the core of these trial-and-error loops lies evaluation: the process of obtaining feedback on candidate solutions via verifiers, simulators, or task-specific scoring functions. While prior work has highlighted the importance of evaluation, it has not explicitly formulated the problem of how evaluation-driven discovery loops can be scaled up in a principled and effective manner to push the boundaries of scientific discovery, a problem this paper seeks to address. We introduce Simple Test-time Evaluation-driven Scaling (SimpleTES), a general framework that strategically combines parallel exploration, feedback-driven refinement, and local selection, revealing substantial gains unlocked by scaling evaluation-driven discovery loops along the right dimensions. Across 21 scientific problems spanning six domains, SimpleTES discovers state-of-the-art solutions using gpt-oss models, consistently outperforming both frontier-model baselines and sophisticated optimization pipelines. Particularly, we sped up the widely used LASSO algorithm by over 2x, designed quantum circuit routing policies that reduce gate overhead by 24.5%, and discovered new Erdos minimum overlap constructions that surpass the best-known results. Beyond novel discoveries, SimpleTES produces trajectory-level histories that naturally supervise feedback-driven learning. When post-trained on successful trajectories, models not only improve efficiency on seen problems but also generalize to unseen problems, discovering solutions that base models fail to uncover. Together, our results establish effective evaluation-driven loop scaling as a central axis for advancing LLM-driven scientific discovery, and provide a simple yet practical framework for realizing these gains.

Related papers

• FrontierOR: Benchmarking LLMs' Capacity for Efficient Algorithm Design in Large-Scale Optimization [61.43300970020897]
  Large language models (LLMs) are increasingly used for optimization modeling and solver-code generation.<n>Existing benchmarks are limited to small or simplified examples far below real-world scale and complexity.<n>We introduce FrontierOR, among the first benchmarks to evaluate LLM-based efficient algorithm design for realistic large-scale optimization problems.
  arXiv  Detail & Related papers  (2026-05-24T20:10:42Z)
• Implementing Robust M-Estimators with Certifiable Factor Graph Optimization [5.293483842121928]
  We develop an adaptive reweighting scheme for factor graph-based estimation in robotics.<n>We show how one can easily implement M-estimators with certifiably correct solvers for inner subproblems.
  arXiv  Detail & Related papers  (2026-03-21T20:17:44Z)
• AlphaResearch: Accelerating New Algorithm Discovery with Language Models [60.502137348923156]
  Large language models have made significant progress in complex but easy-to-verify problems, yet they still struggle with discovering the unknown.<n>We present textbfAlphaResearch, an autonomous research agent designed to discover new algorithms on open-ended problems.
  arXiv  Detail & Related papers  (2025-11-11T18:03:22Z)
• Can Large Language Models Help Experimental Design for Causal Discovery? [94.66802142727883]
  Large Language Model Guided Intervention Targeting (LeGIT) is a robust framework that effectively incorporates LLMs to augment existing numerical approaches for the intervention targeting in causal discovery.<n>LeGIT demonstrates significant improvements and robustness over existing methods and even surpasses humans.
  arXiv  Detail & Related papers  (2025-03-03T03:43:05Z)
• An Early FIRST Reproduction and Improvements to Single-Token Decoding for Fast Listwise Reranking [50.81324768683995]
  FIRST is a novel approach that integrates a learning-to-rank objective and leveraging the logits of only the first generated token. We extend the evaluation of FIRST to the TREC Deep Learning datasets (DL19-22), validating its robustness across diverse domains. Our experiments confirm that fast reranking with single-token logits does not compromise out-of-domain reranking quality.
  arXiv  Detail & Related papers  (2024-11-08T12:08:17Z)
• Routing Arena: A Benchmark Suite for Neural Routing Solvers [8.158770689562672]
  We propose a benchmark suite for Routing Problems that provides a seamless integration of consistent evaluation and the provision of baselines and benchmarks prevalent in the Machine Learning- and Operations Research field. A comprehensive first experimental evaluation demonstrates that the most recent Operations Research solvers generate state-of-the-art results in terms of solution quality and runtime efficiency when it comes to the vehicle routing problem.
  arXiv  Detail & Related papers  (2023-10-06T10:24:33Z)
• Multivariate Time Series Anomaly Detection: Fancy Algorithms and Flawed Evaluation Methodology [2.043517674271996]
  We discuss how a normally good protocol may have weaknesses in the context of MVTS anomaly detection. We propose a simple, yet challenging, baseline based on Principal Components Analysis (PCA) that surprisingly outperforms many recent Deep Learning (DL) based approaches on popular benchmark datasets.
  arXiv  Detail & Related papers  (2023-08-24T20:24:12Z)
• Accelerating Inverse Learning via Intelligent Localization with Exploratory Sampling [1.5976506570992293]
  solving inverse problems is a longstanding challenge in materials and drug discovery. Deep generative models are recently proposed to solve inverse problems. We propose a novel approach (called iPage) to accelerate the inverse learning process.
  arXiv  Detail & Related papers  (2022-12-02T08:00:04Z)
• Neural Improvement Heuristics for Graph Combinatorial Optimization Problems [49.85111302670361]
  We introduce a novel Neural Improvement (NI) model capable of handling graph-based problems where information is encoded in the nodes, edges, or both. The presented model serves as a fundamental component for hill-climbing-based algorithms that guide the selection of neighborhood operations for each.
  arXiv  Detail & Related papers  (2022-06-01T10:35:29Z)
• Large-Scale Sequential Learning for Recommender and Engineering Systems [91.3755431537592]
  In this thesis, we focus on the design of an automatic algorithms that provide personalized ranking by adapting to the current conditions. For the former, we propose novel algorithm called SAROS that take into account both kinds of feedback for learning over the sequence of interactions. The proposed idea of taking into account the neighbour lines shows statistically significant results in comparison with the initial approach for faults detection in power grid.
  arXiv  Detail & Related papers  (2022-05-13T21:09:41Z)
• RL-PGO: Reinforcement Learning-based Planar Pose-Graph Optimization [1.4884785898657995]
  This paper presents a state-of-the-art Deep Reinforcement Learning (DRL) based environment and proposed agent for 2D pose-graph optimization. We demonstrate that the pose-graph optimization problem can be modeled as a partially observable Decision Process and evaluate performance on real-world and synthetic datasets.
  arXiv  Detail & Related papers  (2022-02-26T20:10:14Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.