Mindblown: a blog about philosophy.

Estimating Nonstationary Variables via the Kernel Lasso – We explore the use of the kernel Lasso to estimate the unknown covariance matrix for nonstationary covariance matrices and its estimation using the kernel Lasso. We propose a new algorithm called Kernel Lasso (KL) that takes the unknown covariance matrix as an input and optimizes a Kernel […]

Fault Detection for Wireless Capsule Capsule Wireless Capsule – Although there has been much research on smart sensors embedded with wireless communication networks, few real-world applications involve a wireless communications network. The application of a mobile wireless communication network in a remote environment has the potential to lead to benefits beyond the wireless communication network. […]

Interpolating Structural and Function Complexity of Neural Networks – Current Convolutional Neural Networks (CNNs) have been proven to be very successful methods for semantic classification. However, current CNNs use a very deep network architecture and have difficulty in handling the low-level semantic content. In this work, we show that a deep CNN trained on image […]

Towards a Theory of a Semantic Portal – Deep learning is a powerful tool for solving problems that are difficult to classify, i.e., problems that are impossible to classify or not to classify. In this work, we present a deep learning algorithm for this problem, and propose a novel algorithm for analyzing the data. The […]

Towards an Efficient Programming Model for the Algorithm of the Kohonen Sub-committee of the NDA (Nasir No. 246)41256,Logical Solution to the Problem of Fuzzy Synchronization of Commodity Swaps by the Combination of Non-Linear Functions, – We present a method for the solving of the following two problems: finding a common set of all available binary […]

A Simple and Effective Online Clustering Algorithm Using Approximate Kernel – Recently, it was reported that the accuracy of various types of statistical models (data), such as linear models, regression models, and graph models are affected by a statistical imbalance, when the model being studied is not the same one used by the other. This […]

Nonlinear Models in Probabilistic Topic Models – One of the difficulties in the design of causal networks is the lack of knowledge which can be readily obtained from data sets. The goal of this research is to construct a causal model which directly captures the data flows and enables us to learn how these flows […]

Learning to Rank by Minimising the Ranker – This thesis investigates the problem of estimating the best ranking of a class of objects from the user-item comparisons. The problem is formulated firstly as the task of finding the best item for that category. This task has been extensively explored in the literature. The proposed method […]

Predicting Ratings by Compositional Character Structure – In this paper, we present a class of algorithms to improve the recognition of ratings in social media. Most existing techniques used in this work are based on two main methods. In this work, we show how to apply these two methods in a joint framework. We provide […]

Towards a real-time CNN end-to-end translation – Most of the previous works on the problem of inferring the meaning of phrases in English translations have only provided simple solutions when solving a particular translation problem, or when trying to translate a certain sentence in some languages. This paper proposes a new framework for translating phrases […]