Breakfast seminar: Data Intelligence for Reliable AI and Multi-Agent Systems with Xuan-Son Vu

For breakfast, please sign up no later than October 2nd by using this link: Register for breakfast 

Large language models are increasingly deployed as autonomous agents that plan, call tools, and collaborate with other agents to complete complex tasks. As these systems grow more capable, their reliability depends less on the model itself and more on the quality, protection, and flow of data around it. This talk presents our perspective on data intelligence as the foundation for trustworthy AI and multi-agent systems, drawing on our own work in pseudonymisation, data synthesis, and federated learning, together with recent advances in agent reliability and evaluation.

Xuan-Son Vu (Ph.D.) is an Assistant Professor at Lund University of Technology, the CTO of WASP Media and Language, Sweden, and the Co-founder of DeepTensor AB and AI for Vietnam. He received his Ph.D. degree from Umeå University and his M.Sc. degree in Computer Science from Kyungpook National University in Korea, with a focus on NLP and Machine Learning. His work has been primarily focused on knowledge, both acquiring knowledge from text and multimodal data, and using structured knowledge to power downstream applications. His recent research centers on multimodal learning and data intelligence for reliable AI, including pseudonymisation and synthetic data generation for privacy preserving training and evaluation, federated learning across organisations, and the reliability and robustness of LLM based multi-agent systems. He is a reviewer for journals and conferences including TheWebConf, ECAI, ICDM, PAKDD, SSR, SC2, COSE (Computers & Security), IJCAI, ACL, and TPAMI.