Shai Ben-David

Technion – Israel Institute of Technology

Papers

Total Citations

H-Index

About

Shai Ben-David is a foundational figure in computational learning theory, whose work bridges the gap between abstract mathematical frameworks and practical machine learning. His key research areas include statistical learning theory, domain adaptation, and the geometry of learning problems. Ben-David is perhaps best known for his seminal contributions to understanding the fundamental limits of learning, particularly through his work on the theory of learning from different distributions—a cornerstone of modern domain adaptation. His highly cited papers, such as "Localization vs. Identification of Semi-Algebraic Sets" (1993, 13 citations; 1998, 15 citations), formally investigate the information complexity of locating known objects from random samples, a problem central to computer vision and robotics. By comparing the tasks of identification and localization, he provided rigorous insights into how much data is truly needed to determine an object's position. With thousands of citations across his body of work, Ben-David's research has profoundly influenced both theoretical computer science and applied AI, making him a key thinker for any student or researcher exploring the mathematical underpinnings of learning systems.