Heterogeneous Scientific Foundation Model Collaboration

2026-04-29

AI Summary

Current language-centric agents struggle to effectively process non-linguistic, structured data in scientific domains, limiting their applicability to multimodal scientific tasks. This work proposes the Eywa framework, which for the first time enables collaborative reasoning between large language models and heterogeneous scientific foundation models, overcoming the constraint of relying solely on language as the interaction medium. Through a language-guided reasoning interface, integration of heterogeneous models, and a dynamic planning coordination mechanism, Eywa supports three collaborative architectures—EywaAgent, EywaMAS, and EywaOrchestra—enabling single-agent, multi-agent, and planning-coordination paradigms. Experiments demonstrate that this approach significantly enhances performance on structured scientific tasks across physics, life sciences, and social sciences, while reducing dependence on purely linguistic reasoning.

Abstract

Agentic large language model systems have demonstrated strong capabilities. However, their reliance on language as the universal interface fundamentally limits their applicability to many real-world problems, especially in scientific domains where domain-specific foundation models have been developed to address specialized tasks beyond natural language. In this work, we introduce Eywa, a heterogeneous agentic framework designed to extend language-centric systems to a broader class of scientific foundation models. The key idea of Eywa is to augment domain-specific foundation models with a language-model-based reasoning interface, enabling language models to guide inference over non-linguistic data modalities. This design allows predictive foundation models, which are typically optimized for specialized data and tasks, to participate in higher-level reasoning and decision-making processes within agentic systems. Eywa can serve as a drop-in replacement for a single-agent pipeline (EywaAgent) or be integrated into existing multi-agent systems by replacing traditional agents with specialized agents (EywaMAS). We further investigate a planning-based orchestration framework in which a planner dynamically coordinates traditional agents and Eywa agents to solve complex tasks across heterogeneous data modalities (EywaOrchestra). We evaluate Eywa across a diverse set of scientific domains spanning physical, life, and social sciences. Experimental results demonstrate that Eywa improves performance on tasks involving structured and domain-specific data, while reducing reliance on language-based reasoning through effective collaboration with specialized foundation models.

Problem

• heterogeneous foundation models
• scientific domains
• non-linguistic data
• agentic systems
• domain-specific tasks

Innovation

• heterogeneous foundation models
• agentic systems
• scientific reasoning
• multimodal collaboration
• language-model interface