<ul><li>We are currently facing a surge in digital scientific publications, known as "big literature", creating challenges in understanding literature landscapes through traditional manual reviews.</li><li>Recent large language models (LLMs) are proficient in literature comprehension, but they fall short in providing the desired comprehensive, objective, open, and transparent views for systematic reviews.</li><li>LitChat is introduced as an end-to-end, interactive literature agent that enhances LLM agents with data-driven tools for exploring literature through user queries interpretation, source retrieval, knowledge graph construction, and evidence-based insights generation.</li><li>The effectiveness of LitChat is demonstrated through a case study on AI4Health, showcasing its ability to swiftly navigate users through vast literature landscapes with data-supported insights, which are challenging to obtain using traditional methods.</li></ul>

Conversational Exploration of Literature Landscape with LitChat

Discover more