Tomaso Poggio develops computational models of brain function in order to understand human intelligence and to build intelligent machines that can mimic human performance.
Learning to see
For Tomaso Poggio, intelligence means the ability to learn. The essence of learning is the ability to generalize from specific examples. Newborn babies, for example, know very little about the visual world, yet they quickly learn to recognize thousands of objects around them. To accomplish this, they must learn from relatively few specific examples to recognize general categories such as animals, cars, or houses. They must also learn to correctly identify new examples that they have never seen before.
The ease with which our brains seem to accomplish this task is deceptive. Early attempts to build computer vision systems were largely unsuccessful, and served only to emphasize how difficult the task really is. Rather than designing an artificial system from scratch, Poggio's approach has been to look at the brain for answers.
The brain is known to decode visual information in a stepwise manner. Information is relayed from our eyes to a series of brain areas, each of which processes the information before relaying it to the next stage. Poggio, in collaboration with McGovern Institute investigator James Dicarlo, simulates this process on a computer using neural network models based on the known properties of real neurons. A recent version of Poggio's model has proved highly successful -- after the computer model is trained on a collection of images of natural scenes with or without animals, the computer learns to recognize the presence of animals with a success rate similar to that of human observers. The computer even makes the same mistakes as humans, suggesting that Poggio's model solves the visual task the same way we do.
This simple model uses only "forward" signals. Poggio is now working on more complex models that mimick the forward and feedback signals within the human brain. This work may help explain brain disorders such as schizophrenia and autism, which Poggio suspects may involve problems associated with feedback signals or their integration with forward signals.
Tomaso Poggio is Eugene McDermott Professor in the Department of Brain and Cognitive Sciences and at the Artificial Intelligence Laboratory. He is a founding member of the McGovern Institute, and is also the director of the Center for Brains, Minds, and Machines, a multi-institutional collaboration headquartered at the McGovern Institute. He joined the MIT faculty in 1981, after ten years at the Max Planck Institute for Biology and Cybernetics in Tubingen, Germany. He received a Ph.D. in 1970 from the University of Genoa. Poggio is a Foreign Member of the Italian Academy of Sciences and a Fellow of the American Academy of Arts and Sciences. He was awarded the 2014 Swartz Prize for Theoretical and Computational Neuroscience.