ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models
🤖 AI Summary
This study addresses the low efficiency in identifying innovative scientific questions across disciplinary boundaries. To this end, we propose ResearchAgent—a novel system that (1) enhances cross-domain literature retrieval via an academic knowledge graph and concept-level entity indexing; (2) employs a multi-agent collaborative framework integrating LLM-driven modules for research question generation, methodology design, and experimental planning; and (3) introduces a human-preference-aligned review agent that simulates peer review to iteratively refine proposals. ResearchAgent is the first to deeply integrate academic graphs, concept knowledge bases, and human-feedback-informed prompt engineering. Experimental evaluation on multidisciplinary papers demonstrates that its generated research proposals significantly outperform baselines in novelty, clarity, and feasibility—validated both by domain experts and automated model-based assessment.
Application Category
📝 Abstract
The pace of scientific research, vital for improving human life, is complex, slow, and needs specialized expertise. Meanwhile, novel, impactful research often stems from both a deep understanding of prior work, and a cross-pollination of ideas across domains and fields. To enhance the productivity of researchers, we propose ResearchAgent, which leverages the encyclopedic knowledge and linguistic reasoning capabilities of Large Language Models (LLMs) to assist them in their work. This system automatically defines novel problems, proposes methods and designs experiments, while iteratively refining them based on the feedback from collaborative LLM-powered reviewing agents. Specifically, starting with a core scientific paper, ResearchAgent is augmented not only with relevant publications by connecting information over an academic graph but also entities retrieved from a knowledge store derived from shared underlying concepts mined across numerous papers. Then, mimicking a scientific approach to improving ideas with peer discussions, we leverage multiple LLM-based ReviewingAgents that provide reviews and feedback via iterative revision processes. These reviewing agents are instantiated with human preference-aligned LLMs whose criteria for evaluation are elicited from actual human judgments via LLM prompting. We experimentally validate our ResearchAgent on scientific publications across multiple disciplines, showing its effectiveness in generating novel, clear, and valid ideas based on both human and model-based evaluation results. Our initial foray into AI-mediated scientific research has important implications for the development of future systems aimed at supporting researchers in their ideation and operationalization of novel work.
Problem
Research questions and friction points this paper is trying to address.
Enhance research productivity with AI
Generate novel scientific ideas automatically
Iteratively refine research using LLM feedback
Innovation
Methods, ideas, or system contributions that make the work stand out.
LLMs for idea generation
Iterative feedback with ReviewingAgents
Knowledge integration from academic graph
🔎 Similar Papers
No similar papers found.
