Visual Cortex and Deep Networks

Learning Invariant Representations

Hardcover
$35.00 US
On sale Sep 23, 2016 | 136 Pages | 9780262034722
A mathematical framework that describes learning of invariant representations in the ventral stream, offering both theoretical development and applications.

The ventral visual stream is believed to underlie object recognition in primates. Over the past fifty years, researchers have developed a series of quantitative models that are increasingly faithful to the biological architecture. Recently, deep learning convolution networks—which do not reflect several important features of the ventral stream architecture and physiology—have been trained with extremely large datasets, resulting in model neurons that mimic object recognition but do not explain the nature of the computations carried out in the ventral stream. This book develops a mathematical framework that describes learning of invariant representations of the ventral stream and is particularly relevant to deep convolutional learning networks.

The authors propose a theory based on the hypothesis that the main computational goal of the ventral stream is to compute neural representations of images that are invariant to transformations commonly encountered in the visual environment and are learned from unsupervised experience. They describe a general theoretical framework of a computational theory of invariance (with details and proofs offered in appendixes) and then review the application of the theory to the feedforward path of the ventral stream in the primate visual cortex.

Tomaso A. Poggio is Eugene McDermott Professor in the Department of Brain and Cognitive Sciences at MIT, where he is also Director of the Center for Brains, Minds, and Machines and Codirector of the Center for Biological and Computational Learning. He is coeditor of Perceptual Learning (MIT Press).

Fabio Anselmi is a Postdoctoral Fellow in the Istituto Italiano di Tecnologia Laboratory for Computational and Statistical Learning at MIT and part of the Center for Brains, Minds, and Machines.

About

A mathematical framework that describes learning of invariant representations in the ventral stream, offering both theoretical development and applications.

The ventral visual stream is believed to underlie object recognition in primates. Over the past fifty years, researchers have developed a series of quantitative models that are increasingly faithful to the biological architecture. Recently, deep learning convolution networks—which do not reflect several important features of the ventral stream architecture and physiology—have been trained with extremely large datasets, resulting in model neurons that mimic object recognition but do not explain the nature of the computations carried out in the ventral stream. This book develops a mathematical framework that describes learning of invariant representations of the ventral stream and is particularly relevant to deep convolutional learning networks.

The authors propose a theory based on the hypothesis that the main computational goal of the ventral stream is to compute neural representations of images that are invariant to transformations commonly encountered in the visual environment and are learned from unsupervised experience. They describe a general theoretical framework of a computational theory of invariance (with details and proofs offered in appendixes) and then review the application of the theory to the feedforward path of the ventral stream in the primate visual cortex.

Author

Tomaso A. Poggio is Eugene McDermott Professor in the Department of Brain and Cognitive Sciences at MIT, where he is also Director of the Center for Brains, Minds, and Machines and Codirector of the Center for Biological and Computational Learning. He is coeditor of Perceptual Learning (MIT Press).

Fabio Anselmi is a Postdoctoral Fellow in the Istituto Italiano di Tecnologia Laboratory for Computational and Statistical Learning at MIT and part of the Center for Brains, Minds, and Machines.

National Depression Education and Awareness Month

For National Depression Education and Awareness Month in October, we are sharing a collection of titles that educates and informs on depression, including personal stories from those who have experienced depression and topics that range from causes and symptoms of depression to how to develop coping mechanisms to battle depression.

Read more

Books for LGBTQIA+ History Month

For LGBTQIA+ History Month in October, we’re celebrating the shared history of individuals within the community and the importance of the activists who have fought for their rights and the rights of others. We acknowledge the varying and diverse experiences within the LGBTQIA+ community that have shaped history and have led the way for those

Read more

Books for Latinx & Hispanic Heritage Month

Penguin Random House Education is proud to celebrate Latinx & Hispanic Heritage Month, which runs annually from September 15th through October 15th.  We are highlighting the works of our authors and illustrators from the Latinx and Hispanic community, whose stories and characters have a profound impact on our society. Here is a collection of titles

Read more