Deep Learning Wins Dogs vs Cats competition on Kaggle
A Deep learning expert wins Kaggle Dogs vs Cats image competition with an almost perfect result.
Yann LeCun, a leading researcher on Deep Learning, who was recently hired by Facebook to head their AI Lab, reports that his former student +Pierre Sermanet won the
Dogs vs Cats competition on Kaggle.
Pierre entry was amazingly good - 98.9% correct. He posted on Google+
The second entry by Orchid seems to be DeCaf, which used convnet feature extractor from Berkeley.
This follows Pierre's 1st place in the ImageNet localization competition and his recent post-competition record on the ImageNet detection task.
convnet is a fast C++/CUDA implementation of convolutional (or more generally, feed-forward) neural networks. It can model arbitrary layer connectivity and network depth. Any directed acyclic graph of layers will do. Training is done using the back-propagation algorithm.
Pierre entry was amazingly good - 98.9% correct. He posted on Google+
I just won the Dogs vs. Cats Kaggle competition, using the deep learning library I wrote during my PhD: OverFeat
The second entry by Orchid seems to be DeCaf, which used convnet feature extractor from Berkeley.
This follows Pierre's 1st place in the ImageNet localization competition and his recent post-competition record on the ImageNet detection task.
convnet is a fast C++/CUDA implementation of convolutional (or more generally, feed-forward) neural networks. It can model arbitrary layer connectivity and network depth. Any directed acyclic graph of layers will do. Training is done using the back-propagation algorithm.