Research in the Center for Biological and Computational Learning is focused on the problem of learning in a) theory; b) engineering applications; c) neuroscience.
In computational neuroscience, the lab has developed a model of the ventral stream in visual cortex which accounts for many physiological data (in V1,
V4 and IT) and psychophysical experiments in difficult object recognition tasks (rapid categorization). The model performs well when compared with computer vision systems on standard bench marks. Present work is extending the architecture to deal with recogntion in video sequences and to include top-down, "attentional" projections.
Watch a presentation about Professor Poggio's work
Visions of the Future - Tomaso Poggio, Thomas Serre and Aude Oliva.
* BBC -- This is part of the excellent BBC series entitled "visions of the future". This short clip here shows work performed at CBCL (MIT) about a computational neuroscience model of the ventral stream of the visual cortex. The story here focuses on recent work by Serre, Oliva and Poggio on comparing the performance of the model to human observers during a rapid object categorization task.
"A Canonical Neural Circuit for Cortical Nonlinear Operations" (Kouh, M . and Poggio, T.) June 2008, Vol. 20, No. 6, Pages 1427-1451 Posted Online April 17, 2008. (doi:10.1162/neco.2008.02-07-466) PDF
"A Model of V4 Shape Selectivity and Invariance," (Cadieu, C., M. Kouh, A. Pasupathy, C. Connor, M. Riesenhuber, and T. Poggio) Journal of Neurophysiology, Vol. 98, 1733-1750, June, 2007. PDF
"Biologically Inspired System for Action Recognition," (Jhuang, H., T. Serre, L. Wolf and T. Poggio) In: Proceedings of the Eleventh IEEE International Conference on Computer Vision (ICCV), 1-8, 2007. PDF
"A Feedforward Architecture Accounts for Rapid Categorization," (Serre, T., A. Oliva and T. Poggio) Proceedings of the National Academy of Sciences (PNAS), Vol. 104, No. 15, 6424-6429, 2007. PDF
"Recognition with Cortex-like Mechanisms," (Serre, T., L. Wolf, S. Bileschi, M. Riesenhuber and T. Poggio) IEEE Transactions on Pattern Analysis and Machine Intelligence, 29, 3, 411-426, 2007. PDF
TECHNOLOGY REVIEW by Fred Hapgood (July 11, 2006): Reverse-Engineering the Brain - At MIT, neuroscience and artificial intelligence are beginning to intersect. - Earl Miller, Jim DiCarlo and Tomaso Poggio. PDF
"Fast Readout of Object Identity from Macaque Inferior Temporal Cortex," (Hung, C.P., G. Kreiman, T. Poggio and J.J. DiCarlo). Science, Vol. 310, 863-866, 2005. PDF
"A Theory of Object Recognition: Computations and Circuits in the Feedforward Path of the Ventral Stream in Primate Visual Cortex," (Serre, T., M. Kouh, C. Cadieu, U. Knoblich, G. Kreiman and T. Poggio). CBCL Paper #259/AI Memo #2005-036, Massachusetts Institute of Technology, Cambridge, MA, October, 2005. PDF
"General Conditions for Predictivity in Learning Theory," (Poggio, T., R. Rifkin, S. Mukherjee and P. Niyogi). Nature, Vol. 428, 419-422, 2004. PDF
"Generalization in Vision and Motor Control," (Poggio, T. and E. Bizzi). Nature, Vol. 431, 768-774, 2004. PDF
"The Mathematics of Learning: Dealing with Data," (Poggio, T. and S. Smale). Notices of the AMS, Vol. 50, No. 5, 537-544, May 2003. PDF
"Models of Object Recognition," (Riesenhuber, M. and T. Poggio). Nature Neuroscience, 3 Supp., 1199-1204, 2000. PDF
"A theory of how the brain might work," (T. Poggio). In Cold Spring Harbor Symposia on Quantitative Biology, LV. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 899-910, 1990.