aiqbench analysis
The paper uses a large language model as the optimizer in a cosmological simulation calibration loop, replacing the black-box Bayesian optimization common in this field. The authors report competitive results at a fraction of the CPU cost when the agent is given informal domain priors such as documentation and heuristics. The evidence is a workshop-level proof of concept, so readers should watch whether the gains hold across broader simulation codes and harder parameter spaces.
Agentic Systems for Calibrating Cosmological Simulations
arXiv:2610.40685v1 Announce Type: cross Abstract: Calibrating a cosmological simulation is an expensive, high-dimensional parameter optimization task. Previous works employ Bayesian optimization for automation, which treats the simulator as a black box and precludes the incorporation of human domain knowledge or workflow into the selection policy. We propose AgenticTuner, an agentic calibration pipeline with a present-day large language model serving as the optimizer who proposes the next parameter trial. We show that these agents can achieve competitive calibration performance at a fraction of the CPU cost when they are given informal, human-readable priors (such as codebase, documentation, and heuristic guides that seasoned scientists may possess). We hope this work serves as an early step toward AI scientists that can carry out complex, long-horizon scientific tasks.
Abstract:Calibrating a cosmological simulation is an expensive, high-dimensional parameter optimization task. Previous works employ Bayesian optimization for automation, which treats the simulator as a black box and precludes the incorporation of human domain knowledge or workflow into the selection policy. We propose AgenticTuner, an agentic calibration pipeline with a present-day large language model serving as the optimizer who proposes the next parameter trial. We show that these agents can achieve competitive calibration performance at a fraction of the CPU cost when they are given informal, human-readable priors (such as codebase, documentation, and heuristic guides that seasoned scientists may possess). We hope this work serves as an early step toward AI scientists that can carry out complex, long-horizon scientific tasks.
Accepted to the 9th annual workshop for Machine Learning & the Physical Sciences (ML4PS). A starter code is available on this https URL
Instrumentation and Methods for Astrophysics (astro-ph.IM); Computation and Language (cs.CL)
arXiv:2610.40685 [astro-ph.IM]
(or arXiv:2610.40685v1 [astro-ph.IM] for this version)
https://doi.org/10.48550/arXiv.2610.40685 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 243 words
Related coverage
Field-level weak lensing cosmology with $60$ simulations using multifidelity simulation-based inference
arXiv · Sep 9, 2026
Cross-simulator transfer with foundation model summaries: Towards robust SKA-era reionization inference
arXiv · Aug 28, 2026
Cross-Fitting for Neural Posterior Estimation
arXiv · Oct 21, 2026
Improving the Sensitivity of Gravitational Wave Detection with Weighted Conformal Prediction
arXiv · Sep 11, 2026