Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LLM4SR: A Survey on Large Language Models for Scientific Research

Ziming Luo, Zonglin Yang, Zexin Xu, Wei Yang, Xinya Du

TL;DR

This survey systematically examines how large language models are reshaping scientific research across hypothesis discovery, experiment planning and execution, paper writing, and peer review. It synthesizes task-specific methods, benchmarks, and evaluation frameworks, contrasts LLM-driven approaches with traditional workflows, and highlights current limitations and open challenges. By cataloging major progress and proposing directions for automated validation, reasoning, and human–AI collaboration, the paper aims to guide researchers and practitioners in integrating LLMs into scientific workflows. The work also provides a repository of resources to support adoption and ongoing development in this rapidly evolving field.

Abstract

In recent years, the rapid advancement of Large Language Models (LLMs) has transformed the landscape of scientific research, offering unprecedented support across various stages of the research cycle. This paper presents the first systematic survey dedicated to exploring how LLMs are revolutionizing the scientific research process. We analyze the unique roles LLMs play across four critical stages of research: hypothesis discovery, experiment planning and implementation, scientific writing, and peer reviewing. Our review comprehensively showcases the task-specific methodologies and evaluation benchmarks. By identifying current challenges and proposing future research directions, this survey not only highlights the transformative potential of LLMs, but also aims to inspire and guide researchers and practitioners in leveraging LLMs to advance scientific inquiry. Resources are available at the following repository: https://github.com/du-nlp-lab/LLM4SR

LLM4SR: A Survey on Large Language Models for Scientific Research

TL;DR

This survey systematically examines how large language models are reshaping scientific research across hypothesis discovery, experiment planning and execution, paper writing, and peer review. It synthesizes task-specific methods, benchmarks, and evaluation frameworks, contrasts LLM-driven approaches with traditional workflows, and highlights current limitations and open challenges. By cataloging major progress and proposing directions for automated validation, reasoning, and human–AI collaboration, the paper aims to guide researchers and practitioners in integrating LLMs into scientific workflows. The work also provides a repository of resources to support adoption and ongoing development in this rapidly evolving field.

Abstract

In recent years, the rapid advancement of Large Language Models (LLMs) has transformed the landscape of scientific research, offering unprecedented support across various stages of the research cycle. This paper presents the first systematic survey dedicated to exploring how LLMs are revolutionizing the scientific research process. We analyze the unique roles LLMs play across four critical stages of research: hypothesis discovery, experiment planning and implementation, scientific writing, and peer reviewing. Our review comprehensively showcases the task-specific methodologies and evaluation benchmarks. By identifying current challenges and proposing future research directions, this survey not only highlights the transformative potential of LLMs, but also aims to inspire and guide researchers and practitioners in leveraging LLMs to advance scientific inquiry. Resources are available at the following repository: https://github.com/du-nlp-lab/LLM4SR
Paper Structure (55 sections, 2 figures, 5 tables)

This paper contains 55 sections, 2 figures, 5 tables.

Table of Contents

  1. Introduction
  2. LLMs for Scientific Hypothesis Discovery
  3. Overview
  4. History of Scientific Discovery
  5. Literature-based Discovery
  6. Inductive Reasoning
  7. Development of Methods
  8. Main Trajectory
  9. Inspiration Retrieval Strategy
  10. Feedback Modules
  11. Evolutionary Algorithm
  12. Leveraging Multiple Inspirations
  13. Ranking of Hypotheses
  14. Automatic Research Question Construction
  15. Other Methods
  16. ...and 40 more sections

Figures (2)

  • Figure 1: Schematic overview of the scientific research pipeline covered in this survey. This cyclical process begins with scientific hypothesis discovery, followed by experiment planning and implementation, paper writing, and finally peer reviewing of papers. The experiment planning stage consists of optimizing experiment design and executing research tasks, while the paper writing stage consists of citation text generation, related work generation, and drafting & writing.
  • Figure 2: The main content flow and categorization of this survey.