Improving Scientific Hypothesis Generation with Knowledge Grounded Large Language Models

TL;DR

This work tackles the reliability challenge in hypothesis generation by grounding LLM outputs in external, structured knowledge from domain-specific knowledge graphs. It introduces KG-CoI, a three-module pipeline consisting of KG-guided context retrieval, KG-augmented chain-of-idea generation, and KG-supported hallucination detection, and demonstrates its effectiveness on a newly constructed biomedical hypothesis dataset. Through comprehensive experiments across open- and closed-source LLMs, KG-CoI achieves higher accuracy and F1 scores and reduces hallucinations compared to Direct, CoT, and RAG baselines. The results underscore the value of integrating KGs with LLM reasoning for trustworthy, literature-grounded scientific hypothesis generation with potential real-world impact in biological research.

Abstract

Large language models (LLMs) have demonstrated remarkable capabilities in various scientific domains, from natural language processing to complex problem-solving tasks. Their ability to understand and generate human-like text has opened up new possibilities for advancing scientific research, enabling tasks such as data analysis, literature review, and even experimental design. One of the most promising applications of LLMs in this context is hypothesis generation, where they can identify novel research directions by analyzing existing knowledge. However, despite their potential, LLMs are prone to generating ``hallucinations'', outputs that are plausible-sounding but factually incorrect. Such a problem presents significant challenges in scientific fields that demand rigorous accuracy and verifiability, potentially leading to erroneous or misleading conclusions. To overcome these challenges, we propose KG-CoI (Knowledge Grounded Chain of Ideas), a novel system that enhances LLM hypothesis generation by integrating external, structured knowledge from knowledge graphs (KGs). KG-CoI guides LLMs through a structured reasoning process, organizing their output as a chain of ideas (CoI), and includes a KG-supported module for the detection of hallucinations. With experiments on our newly constructed hypothesis generation dataset, we demonstrate that KG-CoI not only improves the accuracy of LLM-generated hypotheses but also reduces the hallucination in their reasoning chains, highlighting its effectiveness in advancing real-world scientific research.

Figures (8)

• Figure 1: An overview of our proposed KG-CoI for knowledge-grounded hypothesis generation. "KG-R" and "Lit-R" are retrievers for scientific knowledge graphs (KGs) and literature, respectively. "LLM-E", "LLM-G", and "LLM-V" are LLM agents for query enrichment, hypothesis generation, and claim verification, respectively.
• Figure 2: Performance of various methods with different numbers of runs used in the self-consistency setting.
• Figure 3: Prompt template for Direct prompting of LLMs.
• Figure 4: Prompt template for CoT prompting of LLMs.
• Figure 5: Prompt template for RAG prompting of LLMs.