Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Foundation Models for Scientific Discovery: From Paradigm Enhancement to Paradigm Transition

Fan Liu, Jindong Han, Tengfei Lyu, Weijia Zhang, Zhe-Rui Yang, Lu Dai, Cancheng Liu, Hao Liu

TL;DR

Foundation models are argued to catalyze a paradigm shift in science, moving beyond enhanced workflows toward autonomous discovery. The authors introduce a three-stage framework—Meta-Scientific Integration, Hybrid Human-AI Co-Creation, Autonomous Scientific Discovery—to position FMs within experimental, theoretical, computational, and data-driven paradigms and across cross-paradigm pipelines. The paper provides a taxonomy of FM-enabled applications, such as Bayesian-guided experiment design, knowledge-graph–assisted hypothesis generation, neural operators for PDEs, and cross-modal knowledge retrieval, with examples like AlphaFold, GPT-4, PROSE-FD, Latent Neural Operators. It also discusses risks (bias, misinformation, reproducibility, authorship) and sketches a roadmap toward embodied, closed-loop, continual-learning autonomous discovery, urging governance and theoretical grounding.

Abstract

Foundation models (FMs), such as GPT-4 and AlphaFold, are reshaping the landscape of scientific research. Beyond accelerating tasks such as hypothesis generation, experimental design, and result interpretation, they prompt a more fundamental question: Are FMs merely enhancing existing scientific methodologies, or are they redefining the way science is conducted? In this paper, we argue that FMs are catalyzing a transition toward a new scientific paradigm. We introduce a three-stage framework to describe this evolution: (1) Meta-Scientific Integration, where FMs enhance workflows within traditional paradigms; (2) Hybrid Human-AI Co-Creation, where FMs become active collaborators in problem formulation, reasoning, and discovery; and (3) Autonomous Scientific Discovery, where FMs operate as independent agents capable of generating new scientific knowledge with minimal human intervention. Through this lens, we review current applications and emerging capabilities of FMs across existing scientific paradigms. We further identify risks and future directions for FM-enabled scientific discovery. This position paper aims to support the scientific community in understanding the transformative role of FMs and to foster reflection on the future of scientific discovery. Our project is available at https://github.com/usail-hkust/Awesome-Foundation-Models-for-Scientific-Discovery.

Foundation Models for Scientific Discovery: From Paradigm Enhancement to Paradigm Transition

TL;DR

Foundation models are argued to catalyze a paradigm shift in science, moving beyond enhanced workflows toward autonomous discovery. The authors introduce a three-stage framework—Meta-Scientific Integration, Hybrid Human-AI Co-Creation, Autonomous Scientific Discovery—to position FMs within experimental, theoretical, computational, and data-driven paradigms and across cross-paradigm pipelines. The paper provides a taxonomy of FM-enabled applications, such as Bayesian-guided experiment design, knowledge-graph–assisted hypothesis generation, neural operators for PDEs, and cross-modal knowledge retrieval, with examples like AlphaFold, GPT-4, PROSE-FD, Latent Neural Operators. It also discusses risks (bias, misinformation, reproducibility, authorship) and sketches a roadmap toward embodied, closed-loop, continual-learning autonomous discovery, urging governance and theoretical grounding.

Abstract

Foundation models (FMs), such as GPT-4 and AlphaFold, are reshaping the landscape of scientific research. Beyond accelerating tasks such as hypothesis generation, experimental design, and result interpretation, they prompt a more fundamental question: Are FMs merely enhancing existing scientific methodologies, or are they redefining the way science is conducted? In this paper, we argue that FMs are catalyzing a transition toward a new scientific paradigm. We introduce a three-stage framework to describe this evolution: (1) Meta-Scientific Integration, where FMs enhance workflows within traditional paradigms; (2) Hybrid Human-AI Co-Creation, where FMs become active collaborators in problem formulation, reasoning, and discovery; and (3) Autonomous Scientific Discovery, where FMs operate as independent agents capable of generating new scientific knowledge with minimal human intervention. Through this lens, we review current applications and emerging capabilities of FMs across existing scientific paradigms. We further identify risks and future directions for FM-enabled scientific discovery. This position paper aims to support the scientific community in understanding the transformative role of FMs and to foster reflection on the future of scientific discovery. Our project is available at https://github.com/usail-hkust/Awesome-Foundation-Models-for-Scientific-Discovery.
Paper Structure (13 sections, 2 figures, 1 table)

This paper contains 13 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background and Preliminary
  3. Rethinking Scientific Paradigms in the Era of Foundation Models
  4. Foundation Model Integration Across Scientific Paradigms
  5. Experiment-Driven Paradigm
  6. Theory-Driven Scientific Paradigm
  7. Computation-Driven Scientific Paradigm
  8. Data-Driven Scientific Paradigm
  9. Cross-Paradigm Foundation Model Integration
  10. Risks of Emerging FM-Centered Scientific Paradigms and Future Direction
  11. Risks of Emerging FM-driven Scientific Paradigms
  12. Future Directions: Toward Autonomous Scientific Discovery
  13. Conclusions

Figures (2)

  • Figure 1: Evolving scientific paradigms empowered by FMs. FMs progressively transition from tool-like infrastructure (meta-scientific integration), to interactive co-creators (hybrid human–AI collaboration), and ultimately to autonomous agents capable of end-to-end scientific discovery.
  • Figure 2: A roadmap of scientific discovery paradigms and their epistemic capabilities.