Accelerating Next-Generation AI in Healthcare through Knowledge Graphs and LLM Co-Learning | College of Computing and Data Science

Large language models (LLMs) have significantly advanced information technology, particularly in data access and analysis, although they struggle with factual knowledge and reasoning, especially in healthcare contexts. Knowledge graphs (KGs) are vital for organizing and indexing biomedical data but face challenges in scalability due to the complexity and variability of healthcare data. This presentation discusses the integration of LLMs and KGs to improve healthcare applications, highlighting ongoing research into LLM-aided KG construction and KG-guided LLM enhancement. The speaker, Carl Yang, an Assistant Professor at Emory University, emphasizes the potential of KG-LLM co-learning and the need for further exploration within real healthcare settings. Carl’s extensive research background includes over 200 publications and multiple accolades in data mining and biomedical informatics. He advocates for collaborations with researchers in data mining and biomedical informatics to further advance these technologies in healthcare.

Source link