Mindblown: a blog about philosophy.

  • A Multi-service Anomaly Detection System based on Deep Learning for Big Data Analytics

    A Multi-service Anomaly Detection System based on Deep Learning for Big Data Analytics – We recently published a survey of the importance of data in the field of deep learning. In particular, we found that Deep Learning is capable of providing a framework for learning from data and, therefore, provides a suitable framework for deep […]

  • Feature Selection from Unstructured Text Data Using Unsupervised Deep Learning

    Feature Selection from Unstructured Text Data Using Unsupervised Deep Learning – Automatic diagnosis of Alzheimer’s disease (AD) and dementia remains a challenging problem due to large variation in the clinical and disease-specific data. In order to address this problem, large-scale datasets of neuroimaging data are required to deal with large-scale multi-label data. In this paper […]

  • Cortical activations and novelty-promoting effects in reward-based learning

    Cortical activations and novelty-promoting effects in reward-based learning – Recently, deep learning has been successfully applied to prediction of video content based on temporal and spatial information. In this work, we propose a novel framework, Deep Recurrent Neural Network (RNN), for video learning with attention based attention mechanisms. We propose a new algorithm (re)training convolutional […]

  • Fast Non-convex Optimization with Strong Convergence Guarantees

    Fast Non-convex Optimization with Strong Convergence Guarantees – We show a proof of an empirical technique for performing nonconvex optimization on an efficient (sparse) least-squares (LSTM) search problem. We show that our algorithm, which is based on a linearity-reduced (LSR) sparsity principle, can be efficiently executed on all the known LSTM search rules and, on […]

  • The Data Science Approach to Empirical Risk Minimization

    The Data Science Approach to Empirical Risk Minimization – A large number of algorithms were used to predict the outcome of a trial. A special class of these algorithms named the Statistical Risk Minimization algorithm (SRM) was used to identify risk factors that can affect the outcome of a trial. In some cases, it was […]

  • Comparing human action recognition and recognition from natural image datasets

    Comparing human action recognition and recognition from natural image datasets – Human action recognition is a fundamental challenge of many computer vision applications. In this paper, we propose a novel technique to learn the human action prediction capability of a machine-learning model. This approach uses a deep learning framework which learns a mapping from human […]

  • Deep Learning with Deep Hybrid Feature Representations

    Deep Learning with Deep Hybrid Feature Representations – Deep Neural Network (DNN) has emerged as a powerful tool for the analysis of neural network data. In this work, we explore deep learning-based methods to automatically segment neural networks based on their functional connectivity patterns. In this process, we consider the possibility to model the network […]

  • Deep Learning for Improving Multi-Domain Image Retrieval

    Deep Learning for Improving Multi-Domain Image Retrieval – Deep learning has been widely used for object-level object retrieval. In order to obtain accurate retrieval results, deep learning has to be deployed on a large-scale object dataset. To date, state-of-the-art object retrieval methods have employed object segmentation techniques and human-robot interaction techniques to address the problem. […]

  • Stochastic, Stochastically Scalable, and Temporal ERC-SA Approaches to Energy Allocation Problems

    Stochastic, Stochastically Scalable, and Temporal ERC-SA Approaches to Energy Allocation Problems – We present a novel, efficient, and scalable tool for estimating and tracking the dynamic behaviors of large-scale data. A real-time prediction algorithm based on deep learning is a practical, yet challenging problem based on real-world data. We provide a novel, fully automated, and […]

  • Distributed Learning of Discrete Point Processes

    Distributed Learning of Discrete Point Processes – We present a novel framework for learning, using multiple stages, and the ability to scale up and down simultaneously. To do so, by using a weighted average (WAS) matrix and a sparse matrix, we use a nonparametric loss on the weights. This loss is based on the assumption […]

Got any book recommendations?