Timezone: »

The Grammar of Vision: Probabilistic Grammar-Based Models for Visual Scene Understanding and Object Categorization
Virginia Savova · Josh Tenenbaum · Leslie Kaelbling · Alan L Yuille

Sat Dec 08 07:30 AM -- 06:30 PM (PST) @ Westin: Alpine A-C
Event URL: http://web.mit.edu/cocosci/nips07.html »

The human ability to acquire a visual concept from a few examples, and to recognize instances of that concept in the context of a complex scene, poses a central challenge to the fields of computer vision, cognitive science, and machine learning. Representing visual objects and scenes as the human mind does is likely to require structural sophistication, something akin to a grammar for image parsing, with multiple levels of hierarchy and abstraction, rather than the "flat" feature vectors which are standard in most statistical pattern recognition. Grammar-based approaches to vision have been slow to develop, largely due to the absence of effective methods for learning and inference under uncertainty. However, recent advances in machine learning and statistical models for natural language have inspired a renewed interest in structural representations of visual objects, categories, and scenes. The result is a new and emerging body of research in computational visual cognition that combines sophisticated probabilistic methods for learning and inference with classical grammar-based approaches to representation. The goal of our workshop is to explore these new directions, in the context of several interdisciplinary connections that converge distinctively at NIPS. We will focus on these challenges: How can we learn better probabilistic grammars for machine vision by drawing on state-of-the-art methods in statistical machine learning or natural language learning? What can probabilistic grammars for machine vision tell us about human visual cognition? How can human visual cognition inspire new developments in computational vision and machine learning?

Author Information

Virginia Savova (Massachusetts Institute of Technology)
Josh Tenenbaum (MIT)

Josh Tenenbaum is an Associate Professor of Computational Cognitive Science at MIT in the Department of Brain and Cognitive Sciences and the Computer Science and Artificial Intelligence Laboratory (CSAIL). He received his PhD from MIT in 1999, and was an Assistant Professor at Stanford University from 1999 to 2002. He studies learning and inference in humans and machines, with the twin goals of understanding human intelligence in computational terms and bringing computers closer to human capacities. He focuses on problems of inductive generalization from limited data -- learning concepts and word meanings, inferring causal relations or goals -- and learning abstract knowledge that supports these inductive leaps in the form of probabilistic generative models or 'intuitive theories'. He has also developed several novel machine learning methods inspired by human learning and perception, most notably Isomap, an approach to unsupervised learning of nonlinear manifolds in high-dimensional data. He has been Associate Editor for the journal Cognitive Science, has been active on program committees for the CogSci and NIPS conferences, and has co-organized a number of workshops, tutorials and summer schools in human and machine learning. Several of his papers have received outstanding paper awards or best student paper awards at the IEEE Computer Vision and Pattern Recognition (CVPR), NIPS, and Cognitive Science conferences. He is the recipient of the New Investigator Award from the Society for Mathematical Psychology (2005), the Early Investigator Award from the Society of Experimental Psychologists (2007), and the Distinguished Scientific Award for Early Career Contribution to Psychology (in the area of cognition and human learning) from the American Psychological Association (2008).

Leslie Kaelbling (MIT)
Alan L Yuille (UCLA)

More from the Same Authors