Neural Abstractive Summarization: Advances in Long Document Processing and Context Control

Watch a 46-minute research talk exploring neural abstractive summarization techniques presented by Professor Grigorios Tsoumakas at the Finnish Center for Artificial Intelligence. Discover a divide-and-conquer approach for summarizing lengthy documents, particularly scientific articles, and learn about Bayesian active summarization that combines active learning with cutting-edge summarization models. Explore methods for controlling summarization output based on specific contexts like topics, along with corresponding evaluation metrics. Examine practical applications of these techniques in healthcare and finance sectors. The speaker, a distinguished Professor of Machine Learning and Knowledge Discovery at the Aristotle University of Thessaloniki, brings extensive expertise in supervised learning and natural language processing, with notable achievements including the ECML PKDD 10-Year Test of Time Award and over 19,000 research citations.