ESCO Webinar #24: Deep Learning Driven Combinatorial Optimization and its Applications
Speaker: Wen Song, Shandong University, China
Abstract
Combinatorial optimization is one of the core technologies for intelligent decision-making. Traditional solution algorithms rely on expert experience, often entailing high development costs and relatively poor adaptability. Deep learning driven combinatorial optimization methods, leveraging the powerful representation learning capabilities of deep neural networks, aim to automatically acquire problem-specific knowledge from historical data in a data-driven manner to guide the solution process, offering the potential to overcome the limitations of traditional approaches. This talk will introduce the fundamental principles and main paradigms of this emerging research hotspot, along with the speaker's latest advances in the field and some insights into future directions.