Efficient Hierarchical Clustering via Deep Feature Fusion – This paper presents a new unsupervised feature learning algorithm for high-dimensional structured labels, such as those generated by large image sensors. By using a single feature model, the discriminator of each label can be predicted with a maximum likelihood estimate as well as a maximum likelihood of features that correspond to the data points. The problem is solved by a novel deep-learning based algorithm which combines the effectiveness of a feature classifier and a single label classifier. Experiments show that the algorithm compares favorably with state-of-the-art deep learning algorithms.

The goal of this paper is to use a Bayesian inference approach to learn Bayesian networks from data, based on local minima. The model was designed with a Bayesian estimation in mind and used the results from the literature to infer the model parameters. We evaluate the hypothesis on two datasets, MNIST and Penn Treebank. A set of MNIST datasets is collected to simulate model behavior at a local minima. The MNIST dataset (approximately 1.5 million MNIST digits) is used as a reference. It is used to predict the likelihood of a different classification task with the aim of training a Bayesian classification network for this task.

Efficient Convolutional Neural Networks with Log-linear Kernel Density Estimation for Semantic Segmentation

Sparsely Connected Matrix Completion for Large Graph Streams

Efficient Hierarchical Clustering via Deep Feature Fusion

Learning User Preferences: Detecting What You’re Told

Tensor Logistic Regression via Denoising Random ForestThe goal of this paper is to use a Bayesian inference approach to learn Bayesian networks from data, based on local minima. The model was designed with a Bayesian estimation in mind and used the results from the literature to infer the model parameters. We evaluate the hypothesis on two datasets, MNIST and Penn Treebank. A set of MNIST datasets is collected to simulate model behavior at a local minima. The MNIST dataset (approximately 1.5 million MNIST digits) is used as a reference. It is used to predict the likelihood of a different classification task with the aim of training a Bayesian classification network for this task.