AgenticSciML: Collaborative Multi-Agent Systems for Emergent Discovery in Scientific Machine Learning

TL;DR

AgenticSciML introduces a multi-agent system where specialized AI agents collaboratively generate, critique, and refine scientific machine learning solutions, leading to significant performance improvements and novel methodologies in physics-informed and operator learning tasks.

Contribution

This work presents a novel multi-agent framework that enables collaborative reasoning and emergent innovation in SciML, reducing reliance on expert-driven design and extensive experimentation.

Findings

Outperforms single-agent and human-designed baselines by up to four orders of magnitude in error reduction.

Generates novel architectures and strategies not explicitly present in the knowledge base.

Demonstrates emergent methodological innovation through collaborative AI reasoning.

Abstract

Scientific Machine Learning (SciML) integrates data-driven inference with physical modeling to solve complex problems in science and engineering. However, the design of SciML architectures, loss formulations, and training strategies remains an expert-driven research process, requiring extensive experimentation and problem-specific insights. Here we introduce AgenticSciML, a collaborative multi-agent system in which over 10 specialized AI agents collaborate to propose, critique, and refine SciML solutions through structured reasoning and iterative evolution. The framework integrates structured debate, retrieval-augmented method memory, and ensemble-guided evolutionary search, enabling the agents to generate and assess new hypotheses about architectures and optimization procedures. Across physics-informed learning and operator learning tasks, the framework discovers solution methods that outperform single-agent and human-designed baselines by up to four orders of magnitude in error reduction. The agents produce novel strategies -- including adaptive mixture-of-expert architectures, decomposition-based PINNs, and physics-informed operator learning models -- that do not appear explicitly in the curated knowledge base. These results show that collaborative reasoning among AI agents can yield emergent methodological innovation, suggesting a path toward scalable, transparent, and autonomous discovery in scientific computing.