This presentation will explore recent advancements in Decision-Focused Learning (DFL), an emerging approach in artificial intelligence (AI) that combines prediction and optimization to train models for optimal decision-making. DFL promises to transform decision-making in various real-world applications by addressing the challenge of estimating unknown parameters in optimization problems in a novel way. This talk will provide an in-depth analysis of various techniques developed to integrate machine learning and optimization models and introduce a taxonomy of DFL methods distinguished by their unique characteristics. The presentation will divide these techniques into two broad classes: gradient-based and gradient-free methods, with gradient-based techniques further classified into four subclasses. Finally, the presentation will conclude by offering valuable insights into current trends and potential future directions in DFL research.

Bio:   Jayanta Mandi is currently serving as a postdoctoral researcher at the DTAI Research Unit in the Department of Computer Science at KU Leuven. He completed his Ph.D. at Vrije Universiteit Brussel. His research focuses on the integration of constrained optimization and machine learning, aiming to apply artificial intelligence (AI) in real-world scenarios for data-driven decision-making. 

See all upcoming talks at https://www.anl.gov/mcs/lans-seminars