P
Patrice Y. Simard
Researcher at Microsoft
Publications - 143
Citations - 21072
Patrice Y. Simard is an academic researcher from Microsoft. The author has contributed to research in topics: Artificial neural network & Convolutional neural network. The author has an hindex of 47, co-authored 143 publications receiving 18416 citations. Previous affiliations of Patrice Y. Simard include Bell Labs & AT&T.
Papers
More filters
Journal ArticleDOI
Learning long-term dependencies with gradient descent is difficult
TL;DR: This work shows why gradient based learning algorithms face an increasingly difficult problem as the duration of the dependencies to be captured increases, and exposes a trade-off between efficient learning by gradient descent and latching on information for long periods.
Proceedings ArticleDOI
Best practices for convolutional neural networks applied to visual document analysis
TL;DR: A set of concrete bestpractices that document analysis researchers can use to get good results with neural networks, including a simple "do-it-yourself" implementation of convolution with a flexible architecture suitable for many visual document problems.
Proceedings ArticleDOI
Comparison of classifier methods: a case study in handwritten digit recognition
Léon Bottou,Corinna Cortes,Corinna Cortes,John S. Denker,John S. Denker,Harris Drucker,Harris Drucker,Isabelle Guyon,Lawrence D. Jackel,Yann LeCun,U.A. Muller,E. Sackinger,Patrice Y. Simard,Patrice Y. Simard,Vladimir Vapnik +14 more
TL;DR: This paper compares the performance of several classifier algorithms on a standard database of handwritten digits by considering not only raw accuracy, but also training time, recognition time, and memory requirements.
Comparison of learning algorithms for handwritten digit recognition
Yann LeCun,Lawrence D. Jackel,Léon Bottou,Léon Bottou,A. Brunot,Corinna Cortes,Corinna Cortes,John S. Denker,John S. Denker,Harris Drucker,Harris Drucker,Isabelle Guyon,Urs A. Muller,E. Sackinger,Patrice Y. Simard,Patrice Y. Simard,Vladimir Vapnik +16 more
TL;DR: This comparison of several learning algorithms for handwritten digits considers not only raw accuracy, but also rejection, training time, recognition time, and memory requirements.
Journal ArticleDOI
Time Is of the Essence: A Conjecture that Hemispheric Specialization Arises from Interhemispheric Conduction Delay
TL;DR: It is suggested that the large brains of mammals such as elephants and cetaceans will also manifest a high degree of hemispheric specialization if the neural apparatus necessary to perform each high-resolution, time-critical task is gathered in one hemisphere.