Chapter VIІ. Accelerating scientific hypothesis testing through multi-model ai reasoning: case studies in optics

Komarov, Alexis (orcid.org/0009-0009-9544-1725) (2025) Chapter VIІ. Accelerating scientific hypothesis testing through multi-model ai reasoning: case studies in optics . ФОП Ямчинський О.В., м. Київ, Україна, pp. 94-104. ISBN 978-617-8830-09-0

Text CHAPTER VIІ..pdf Download (809kB)

Abstract

This study explores the transformative role of artificial intelligence (AI) tools in accelerating scientific hypothesis validation, focusing on the iterative use of reasoning and deep research large language models (LLMs). We classify LLMs into single-step, multi-step reasoning, and web-integrated deep research variants, demonstrating that ensembles of independent models enhance accuracy probabilistically from 8-27% for individual reasoning models to 48-62% for collective inference, when addressing complex physical queries. To illustrate the approach of collective AI inference, we examine case studies of diamond Bragg mirror reflectivity for keV-scale X-rays and metal vapor lasing for high-boiling-temperature metals. By iteratively prompting six LLMs with tailored queries and peer-reviewed data, we derive optimized theoretical results and experimental setups. The results underscore the ability of LLMs to accelerate scientific hypothesis testing, identify theoretical limits, and design experimental configurations while also highlighting the importance of verifying AI outputs, confronting AI with facts and follow-up questions, and accounting for AI model correlations. This framework pioneers a paradigm shift in interdisciplinary research, merging AI-driven reasoning with domain-specific expertise to resolve ambiguities in cutting-edge material science and photonics.

Item Type:	Book
Keywords:	Artificial intelligence, Reasoning models, Deep research, Diamond Bragg mirrors, Metal vapor lasers
Subjects:	Science and knowledge. Organization. Computer science. Information. Documentation. Librarianship. Institutions. Publications > 00 Prolegomena. Fundamentals of knowledge and culture. Propaedeutics > 004 Computer science and technology. Computing. Data processing > 004.9 Application-oriented computer-based techniques Science and knowledge. Organization. Computer science. Information. Documentation. Librarianship. Institutions. Publications > 5 Мathematics. natural sciences > 53 Physics
Divisions:	Institute for Digitalisation of Education > Generic resouse
Depositing User:	д.пед.н. Анна Яцишин
Date Deposited:	19 Jan 2026 14:54
Last Modified:	19 Jan 2026 14:54
URI:	https://lib.iitta.gov.ua/id/eprint/748238

Downloads

Actions (login required)

View Item