AI for Information Retrieval
We study information retrieval and recommendation with knowledge graphs, large language models, and multimodal data, including bias correction, reliability, and personalization.
AI for Information Retrieval
Members: Wonhyung Shin, JaeHwan Jung, 김지혜, 이수형
Methods: Causal Inference, Recommendation Algorithm, Query Performance Prediction, Task Oriented Dialogue System, Fact Verification
Information retrieval and recommendation have long been central topics in AI, data mining, databases, and machine learning. More recently, their combination with knowledge graphs, large language models (LLMs), and multimodal data has rapidly expanded research into retrieval and recommendation in new environments.
Important challenges include removing popularity effects, correcting biases in data and algorithms, evaluating the factuality and reliability of recommendations, and hyper-personalization. Beyond user–item relationships, there is a growing need to integrate temporal and spatial information and diverse external knowledge as knowledge graphs to analyze and use much more complex relationships.
Understanding and using these complex information structures requires deeper and more scientific analytical methods. We use mathematical modeling based on machine learning to analyze the principles of information retrieval and recommender systems and study new methods to improve accuracy, reliability, and personalization.
Funding: Ministry of Science and ICT
Leader: Wonhyung Shin