Tensor-based transfer learning for image recognition – In this paper, we tackle the problem of classification of multi-class multi-task learning. We first propose a classification method to perform classification in the presence of unseen classes. Then we extend it to handle cases with nonlinearities in the classification problem. The proposed classification method is based on an optimization of the model parameters and our main goal is to provide efficient learning methods for classification. The main result of our method is to improve the classification performance on the large dataset collected using ImageNet. Furthermore, we use an adaptive hashing algorithm to estimate the hash function parameters with a minimal loss and we use it to reduce the variance. Experimental results demonstrate the effectiveness of our approach.

In this paper, we present a new framework for speech understanding in natural language, based on the use of a deep neural network (DNN) to recognize speech phrases. The system first learns a sequence of words to encode the phrase into a vector space using a multi-level feature representation. Next, it uses a neural network to capture the semantic similarity between words, based on the word embedding space and their relation to sentence descriptions. A DNN trained on the word embedding space can recognize both sentences and phrases with higher precision than that provided for by state-of-the-art deep learning methods. Finally, we use these system to develop and test a speech recognition system able to recognize phrases like I’m just a human and I speak English and This is a question. The evaluation of the system shows that it correctly identifies more than 90% of phrases with positive speech-related annotations.

Predicting Daily Activity with a Deep Neural Network

The Role of Intensive Regression in Learning to Play StarCraft

Tensor-based transfer learning for image recognition

High quality structured output learning using single-step gradient discriminant analysis

Predicting Speech Ambiguity of Linguistic Contexts with Multi-Tensor NetworksIn this paper, we present a new framework for speech understanding in natural language, based on the use of a deep neural network (DNN) to recognize speech phrases. The system first learns a sequence of words to encode the phrase into a vector space using a multi-level feature representation. Next, it uses a neural network to capture the semantic similarity between words, based on the word embedding space and their relation to sentence descriptions. A DNN trained on the word embedding space can recognize both sentences and phrases with higher precision than that provided for by state-of-the-art deep learning methods. Finally, we use these system to develop and test a speech recognition system able to recognize phrases like I’m just a human and I speak English and This is a question. The evaluation of the system shows that it correctly identifies more than 90% of phrases with positive speech-related annotations.