Mindblown: a blog about philosophy.

Cascaded Multi-level Feature Space for Graph Embedding – In this paper, a new deep convolutional neural network architecture was proposed, that has the ability to efficiently solve a number of problems. The architecture, that is, a neural network for a hierarchical multi-dimensional manifold space, can cope with multiple multilinear models in a supervised learning scenario. […]

Deep Convolutional Neural Network for Brain Decoding – The recent advances in deep learning in the field of brain decoding have enabled a dramatic reduction in the amount of data that need to be transmitted across the brain to be processed by the neural networks. At this time, the use of convolutions has become a […]

Evolving Minimax Functions via Stochastic Convergence Theory – We propose a general method for estimating the performance of a linear classifier, by using a single, weighted, random sample-based, linear ensemble estimator. Our method has the following advantages: (1) It is equivalent to a weighted Gaussian process; (2) It is robust to any non-linearity; and (3) […]

Probabilistic Models for Time-Varying Probabilistic Inference – We present a machine learning approach to image classification, which utilizes sparse representations for classification tasks. We build on the recent success of unsupervised learning for deep learning, where supervised learning is used to automatically infer an image from some labels. Despite the huge successes of unsupervised training […]

What Language does your model teach you best? – We will use the standard dataset of English spoken by 14,000 people to study the human ability to communicate verbally. To learn and predict these sentences, we use a deep learning model called Machine-Net – which has been trained to predict words and phrases. It was […]

An Open Source Framework for Video Processing from Natural Scene Data – In this paper, we propose a new approach for extracting visual concepts from the observed scene. We first extract the scene features, and then use a deep neural network to extract the semantic features. The proposed approach is based on minimizing the variance […]

A Hierarchical Latent Model for Learning Distribution Regression – This paper presents a new data-driven method for learning a novel model of human behavior. In particular, it is the model, the model and the model parameters of a learning process. In the first part, the model is composed of a set of different features which […]

Automatic Matching of Naturalistic Images using the Local Frequency Distribution – We present a method for multi-task retrieval that is simple yet effective. We propose to use the input space of visual images to perform an inference layer in the context of the visual search task. Using these inputs, the task is to map these […]

Modelling linguistic discourse structure – We present a method of automatically estimating natural language dialogue systems from data. Using our model, we have obtained results on a wide range of natural language dialogue systems and show that it is possible to estimate the most effective natural language dialogue systems in some scenarios (for example, when […]

Predicting the behavior of interacting nonverbal children through a self-supervised learning procedure – In this paper, we propose a framework for developing a visual classification system that can learn the visual features and labels of a toy, as well as their attributes to the toy. Our framework consists of three stages. First, we formulate the […]