Antonio Torralba

Antonio Torralba is the Delta Electronics Professor of Electrical Engineering and Computer Science at MIT, leading the Artificial Intelligence + Decision-Making (AI+D) faculty within the EECS department. He is affiliated with the Computer Science and Artificial Intelligence Laboratory (CSAIL) and has held roles such as director of the MIT Quest for Intelligence and MIT IBM Watson AI Lab. His research focuses on computer vision, machine learning, and human visual perception, with interests in multimodal systems, neural networks, and computational photography. Torralba has advised numerous PhD and master’s students, including notable alumni such as Tongzhou Wang, Carl Vondrick, and Bolei Zhou. He received a doctoral degree from the Universitat Politècnica de Catalunya (UPC) in 1994 and earned an Honoris Causa from UPC in 2020. His work spans foundational contributions to vision algorithms, generative models, and interpretable AI, with over 300 publications and co-authorship of the textbook Foundations of Computer Vision.

• Education: Doctorate in Electrical Engineering (UPC, 1994)
• Key Roles: Director of MIT Quest for Intelligence (2018–2020), Head of AI+D Faculty (2020–present)
• Research Labs: CSAIL, MIT-IBM Watson AI Lab

His research emphasizes building systems that mimic human perception across modalities (vision, sound, touch) and explores topics like adversarial robustness, multimodal reasoning, and synthetic data generation. Notable projects include the ADE20K dataset, the SUN database, and contributions to neural scene representation learning. Torralba’s work bridges AI theory and practical applications, influencing both academia and industry.

Scientific achievements include pioneering work on context-aware object detection, scene understanding, and generative models. His lab’s contributions span visual question answering, audio-visual reasoning, and embodied AI, with datasets like HACS and CLEVRER advancing benchmarking in action prediction and physical reasoning.

Advising highlights include mentoring over 50 students and postdocs, many of whom have become leading researchers in academia and tech companies. Key grants and initiatives include the MIT Quest for Intelligence, which unites machine learning and decision-making research across EECS.

His artistic projects, such as Average World and Visual Dictionary, visualize large-scale image patterns, reflecting his interest in the intersection of computation and art.