analiza aiqbench
This paper proposes using LLM agents to search for scientific machine learning surrogates while auditing each candidate against fixed physical contracts, such as causality and boundary conditions, separately from validation error. Its central finding is that surrogates chosen on error alone can violate physics even when their error is low, which matters for anyone deploying learned models in engineering or simulation. The results are limited to four solid-mechanics benchmarks, so the evidence supports the audit approach as a promising pattern rather than a general solution.
Physics-Audited Agentic Discovery in Scientific Machine Learning
arXiv:2607.07379v2 Announce Type: replace-cross Abstract: Agentic discovery can yield mechanics surrogates with low validation error that nonetheless violate physical requirements. We present Physics-Audited Agentic SciML (PA-SciML), a framework combining large language model agents with physics auditing. Physical requirements, encoded as contracts such as superposition, boundary conditions, and causality, are fixed before candidate search and tested on predicted fields without reference solutions, including through an adversarial search. Audit results guide later proposals and model selection but are reported separately from validation error. Single-edit tests document reusable design changes. Four solid-mechanics problems of increasing complexity, from parametric elasticity to heterogeneous stiffness fields and cyclic plasticity with combined hardening, reveal causality, zero-input, and stress-bound violations in surrogates selected by error alone. In elastodynamics, two surrogates reach nearly equal error, but only the audit-enabled one is causal by construction. Audit-enabled surrogates pass the required checks at sampled inputs while matching or exceeding audit-disabled accuracy in most comparisons.
Abstract:Agentic discovery can yield mechanics surrogates with low validation error that nonetheless violate physical requirements. We present Physics-Audited Agentic SciML (PA-SciML), a framework combining large language model agents with physics auditing. Physical requirements, encoded as contracts such as superposition, boundary conditions, and causality, are fixed before candidate search and tested on predicted fields without reference solutions, including through an adversarial search. Audit results guide later proposals and model selection but are reported separately from validation error. Single-edit tests document reusable design changes. Four solid-mechanics problems of increasing complexity, from parametric elasticity to heterogeneous stiffness fields and cyclic plasticity with combined hardening, reveal causality, zero-input, and stress-bound violations in surrogates selected by error alone. In elastodynamics, two surrogates reach nearly equal error, but only the audit-enabled one is causal by construction. Audit-enabled surrogates pass the required checks at sampled inputs while matching or exceeding audit-disabled accuracy in most comparisons.
Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
arXiv:2607.07379 [cs.AI]
(or arXiv:2607.07379v2 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2607.07379 Focus to learn more arXiv-issued DOI via DataCite
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.
Text integral sub licența deschisă a publicației · 243 cuvinte