Deep Convolutional Auto-Encoder: Learning Unsophisticated Image Generators from Noisy Labels – We present a new and important technique for image denoising. Specifically, we employ the Convolutional Neural Network to learn to extract image labels from the input data. In order to generate a label to extract the labeling from the input image vector, an algorithm is implemented using a deep convolutional neural network. We perform experiments on the standard datasets of MNIST, SUN, and CIFAR-10. We show that the proposed method significantly outperforms the state-of-the-art methods for denoising performance in all datasets.

We propose a new approach for supervised clustering, where a cluster of nodes is sampled from a random distribution, and a low probability distribution is modeled. The low probability distribution is the subset of the sample which contains all nodes that are sampled from the distribution. An efficient low-rank projection procedure is proposed for this problem. In particular, the projection is formulated as a sub-weight function for the high-dimensional feature representation, which is then used to construct a sparse projection. We first show that the sparse projection is a regularizer for this problem, which, in turn, allows to automatically handle outliers. Second, we show how we can use high-dimensional features represented by such sparse projections to estimate high-dimensional features corresponding to high-dimensional data. Third, we show some practical applications using our approach. We report the proposed process and some results of the implementation of the method for clustering patients with diabetes.

Multi-Instance Dictionary Learning in the Matrix Space with Applications to Video Classification

Extended Version – Probability of Beliefs in Partial-Tracked Bayesian Systems

# Deep Convolutional Auto-Encoder: Learning Unsophisticated Image Generators from Noisy Labels

Towards Effective Deep-Learning Datasets for Autonomous Problem Solving

A Novel Low-Rank Minimization Approach For Clustering Large-Scale Health Data Using A Novel Kernel Ridge Regression ModelWe propose a new approach for supervised clustering, where a cluster of nodes is sampled from a random distribution, and a low probability distribution is modeled. The low probability distribution is the subset of the sample which contains all nodes that are sampled from the distribution. An efficient low-rank projection procedure is proposed for this problem. In particular, the projection is formulated as a sub-weight function for the high-dimensional feature representation, which is then used to construct a sparse projection. We first show that the sparse projection is a regularizer for this problem, which, in turn, allows to automatically handle outliers. Second, we show how we can use high-dimensional features represented by such sparse projections to estimate high-dimensional features corresponding to high-dimensional data. Third, we show some practical applications using our approach. We report the proposed process and some results of the implementation of the method for clustering patients with diabetes.