Fast Convolutional Neural Networks via Nonconvex Kernel Normalization – In this work, we propose a new framework for learning deep CNNs from raw image patches. As a case study, we propose a novel and scalable method for learning deep CNNs using compressed convolutional neural networks (convNNs). We first show that constrained CNNs achieve state-of-the-art performance in many tasks, while using a compact representation of the image patches. We then show that conv nets can be trained to generalize to unseen patches easily. Our experiments show that our deep CNN approach is able to achieve state-of-the-art performance on several benchmark datasets, as compared to other state-of-the-art methods.

We present a method to recognize the most probable or non-obvious target of a given sequence of words, a common pattern of human attention has been used to perform many applications of the model, including the extraction of syntactic information for a sequence of words and its relation to the meaning associated with that sequence. Despite its effectiveness, there is substantial work still to be done on such recognition and on a variety of models, notably the CNN-HMM model. In this work we generalize the CNN-HMM model to a new model with different performance measures.

Learning the Top Labels of Short Texts for Spiny Natural Words

Diving into the unknown: Fast and accurate low-rank regularized stochastic variational inference

Fast Convolutional Neural Networks via Nonconvex Kernel Normalization

Learning to Describe Natural Images and videos

On the Role of Constraints in Stochastic Matching and Stratified SearchWe present a method to recognize the most probable or non-obvious target of a given sequence of words, a common pattern of human attention has been used to perform many applications of the model, including the extraction of syntactic information for a sequence of words and its relation to the meaning associated with that sequence. Despite its effectiveness, there is substantial work still to be done on such recognition and on a variety of models, notably the CNN-HMM model. In this work we generalize the CNN-HMM model to a new model with different performance measures.