Category: Uncategorized

  • Using the Multi-dimensional Bilateral Distribution for Textual Discrimination

    Using the Multi-dimensional Bilateral Distribution for Textual Discrimination – We present a new dataset for a novel kind of semantic discrimination (tense) task aiming at comparing two types of text: semantic and unsemantically. It includes large-scale annotated hand annotated datasets that are large in size and are capable of covering an entire language. We propose […]

  • Segmentation from High Dimensional Data using Gaussian Process Network Lasso

    Segmentation from High Dimensional Data using Gaussian Process Network Lasso – The recent release of Convolutional Neural Networks (CNN) with deep architectures can be easily implemented, but is computationally expensive to train. Recent work has shown that the amount of data needed for training CNNs can be increased with the number of parameters used by […]

  • Efficient Online Convex Sparse Autoencoders for Nonconvex Sparsity

    Efficient Online Convex Sparse Autoencoders for Nonconvex Sparsity – In this paper, we propose a family of online computationally optimal (OFA) inference algorithms for nonconvex and polynomial problems. This algorithm was designed to solve the problem where the data are sparse and contain unknown entries, i.e., those data that are never seen. The algorithm is […]

  • Compositional Argumentation with Inter-rater Agreement

    Compositional Argumentation with Inter-rater Agreement – We present an online framework that generalizes the Markov Decision Process (MDP) to an online environment where we learn to use the inputs and evaluate their performance. The goal is to predict the response of the agent on each of the two inputs to the agent. Our framework, the […]

  • Pose Flow Estimation: Interpretable Interpretable Feature Learning

    Pose Flow Estimation: Interpretable Interpretable Feature Learning – Recent advances in deep learning have enabled the efficient training of deep neural networks, but the large number of datasets still requires a dedicated optimization. To address this problem, it is important for both the training and optimization steps to be made parallel. In this paper, we […]

  • The Entire Model Is Approximately Truncated: An Optimal Estimation of Linear Parameters

    The Entire Model Is Approximately Truncated: An Optimal Estimation of Linear Parameters – This paper proposes the use of a random-assignment algorithm for a multi-modal machine learning problem with a simple but general purpose function. The model must be one of binary data structures, such as a binary tree, or some similar structure, such as […]

  • Using Generalized Cross-Domain-Universal Representations for Topic Modeling

    Using Generalized Cross-Domain-Universal Representations for Topic Modeling – A new paradigm for multi-class classification from a large range of visual cues is proposed, which utilizes the multi-class feature set to guide the classification process. The proposed framework generalizes to a new set of multi-class classes, i.e., an image with more than 6 classes. The proposed […]

  • Distributed Stochastic Dictionary Learning

    Distributed Stochastic Dictionary Learning – This paper proposes a novel stochastic classification framework for binary recognition problems such as classification, clustering, and ranking. Under such models, in order to model uncertainty, one can choose to model the gradient as a mixture of two-valued parameters (i.e., the distance between the output and the input). Here, the […]

  • Fast and Robust Proximal Algorithms for Graph-Structured Variational Computation

    Fast and Robust Proximal Algorithms for Graph-Structured Variational Computation – We present the first-ever model-free stochastic algorithm for the purpose of estimating the likelihood of a target variable, using a combination of two-dimensional probabilistic models. Unlike existing stochastic optimization algorithms that model stochastic processes, our algorithm can also model uncertainty in the underlying stochastic process. […]

  • Relevance Estimation Using Sparse Multidimensional Scaling: Application to Classification and Regression

    Relevance Estimation Using Sparse Multidimensional Scaling: Application to Classification and Regression – We present a method to use unsupervised feature learning (similar to Sparse Multi-Class Classification) over large class images (e.g., MNIST and CIFAR-10). Under certain assumptions about the image representations, we establish a new classifier for the task of classification over the MNIST dataset. […]