Debayan Banerjee 

Date/Time: 27.02.2025, 11:00 

Location: D2.2.094 

Abstract 

LLMs, especially instruction tuned LLMs, have taken the world by storm. Yet, their magical ability is confined to within its own weights. When interacting with external sources of data, old school approaches such as Information Retrieval have to be relied on. Knowledge Graphs represent an even greater challenge, since the structure is not rows and columns, but nodes and edges. What are the current state of the art methods to enable LLMs and KGs to talk to each other, and what does the future hold?

Bio 

Dr. Debayan Banerjee is a postdoctoral researcher at the Leuphana University Lüneburg in the area of Knowledge Graph Question Answering, with his current focus on LLMs. He defended his PhD thesis in August 2024 from University of Hamburg, Germany. He joins WU Vienna and the BILAI cluster as Assistant Professor (postdoc) in April 2025.