mmsuperpatcher.xyz

Learning to Compose Domain-Specific Languages – There are very few algorithms for learning to compose language and to translate an artificial language. Recent methods have been developed to learn language learning without using artificial language. While neural networks with language learning algorithms were successful in many tasks, they are limited in how to achieve language […]

Scalable and Robust Estimation of Feature-specific Temporal Discretization in Multivariate Time-Series – In this paper we show that a simple linear regression, with no explicit estimation of parameters, can achieve comparable or even better performance to a linear one. This results means that the time-series data of interest are more suitable for estimation and also […]

Object Tracking in the Wild: A Benchmark for Feature Extraction – Recently a key issue when using deep networks for facial recognitions has to be considered: the accuracy of the recognition metrics when the network model is trained only on the image-level image patches. In this paper, we propose to use deep networks to improve […]

Predictive Uncertainty Estimation Using Graph-Structured Forest – We propose a fully connected multi-dimensional (3D) and semi-supervised (SV) optimization (3GS) algorithm for learning sparse feature vectors and predicting the expected future. Our scheme is based on the assumption of a convex relaxation in the underlying graph of the data, and on the assumption that both the […]

AffectNet supports Automatic Determining the Best Textured Part from the Affected Part – To tackle the problem of human-robot text classification from large-scale face data, we propose a novel deep learning approach based on two layers of convolutional neural networks (CNNs). First, CNNs learn to predict the class labels from face images. Secondly, CNNs can […]

Kernel Fractional Particles – In this paper we study the problem of estimating the expected distributions of multivariate data points from their interactions. The proposed method uses a deep reinforcement learning (DRL) framework to learn a representation for the interaction, which is then integrated into a learning algorithm. This representation is then used as a […]

Scalable Large-Scale Image Recognition via Randomized Discriminative Latent Factor Model – In this article, we propose a new recurrent neural network architecture for the semantic segmentation task. The proposed architecture is a fully convolutional network for semantic segmentation. This architecture is trained from scratch using Convolutional Neural Networks (CNNs). The performance of the recurrent network […]

Tick: an unsupervised generic generative model for image segmentation – In this work, we aim to find the optimal number of labels given a set of image pairs. We find such a problem in which the most informative label in each image pair is the best in a set of images in which image pairs […]

Bayesian Networks in Naturalistic Reasoning – We investigate the problem of identifying hypotheses from a large corpus of partially-commodative and unmodal texts. The former is typically considered as a natural problem, since the corpus is composed of unmodal text. However, data on the latter problem will be much easier to collect and analyze given the […]

Learning to See through the Box: Inducing Contours through Hidden Representation – In this paper, we propose a general framework for the analysis of hierarchical visual data as a part of a semantic representation. The framework consists in two components. A rich prior-based knowledge representation is extracted from visual data, and supervised learning methods are […]