The State University of New York

United States of America As Represented by the Secretary of the Air Force

Other

The Research Foundation for the State University

Zhongfei Zhang

Bo Long

Zhen Guo

Ruofei Zhang

Jian Yao

John J Salerno

Shuangfei Zhai

Yi Xu

Rohini K Srihari

The general problem of pattern change discovery between high-dimensional data sets is addressed by considering the notion of the principal angles between the subspaces is introduced to measure the subspace difference between two high-dimensional data sets. Current methods either mainly focus on magnitude change detection of low-dimensional data sets or are under supervised frameworks. Principal angles bear a property to isolate subspace change from the magnitude change. To address the challenge of directly computing the principal angles, matrix factorization is used to serve as a statistical framework and develop the principle of the dominant subspace mapping to transfer the principal angle based detection to a matrix factorization problem. Matrix factorization can be naturally embedded into the likelihood ratio test based on the linear models. The method may be unsupervised and addresses the statistical significance of the pattern changes between high-dimensional data sets.

Pattern change discovery between high dimensional data sets

A method of modelling data, comprising: training an objective function of a linear classifier, based on a set of labeled data, to derive a set of classifier weights; defining a posterior probability distribution on the set of classifier weights of the linear classifier; approximating a marginalized loss function for an autoencoder as a Bregman divergence, based on the posterior probability distribution on the set of classifier weights learned from the linear classifier; and classifying unlabeled data using the autoencoder according to the marginalized loss function.

Semisupervised autoencoder for sentiment analysis

Systems and Methods for multi-modal or multimedia image retrieval are provided. Automatic image annotation is achieved based on a probabilistic semantic model in which visual features and textual words are connected via a hidden layer comprising the semantic concepts to be discovered, to explicitly exploit the synergy between the two modalities. The association of visual features and textual words is determined in a Bayesian framework to provide confidence of the association. A hidden concept layer which connects the visual feature(s) and the words is discovered by fitting a generative model to the training image and annotation words. An Expectation-Maximization (EM) based iterative learning procedure determines the conditional probabilities of the visual features and the textual words given a hidden concept class. Based on the discovered hidden concept layer and the corresponding conditional probabilities, the image annotation and the text-to-image retrieval are performed using the Bayesian framework.

System and method for multimedia ranking and multi-modal image retrieval using probabilistic semantic models and expectation-maximization (EM) learning

Multimodal data mining in a multimedia database is addressed as a structured prediction problem, wherein mapping from input to the structured and interdependent output variables is learned. A system and method for multimodal data mining is provided, comprising defining a multimodal data set comprising image information; representing image information of a data object as a set of feature vectors in a feature space; clustering in the feature space to group similar features; associating a non-image representation with a respective image data object based on the clustering; determining a joint feature representation of a respective data object as a mathematical weighted combination of a set of components of the joint feature representation; optimizing a weighting for a plurality of components of the mathematical weighted combination with respect to a prediction error between a predicted classification and a training classification; and employing the mathematical weighted combination for automatically classifying a new data object.

Enhanced max margin learning on multimodal data mining in a multimedia database

Relational clustering has attracted more and more attention due to its phenomenal impact in various important applications which involve multi-type interrelated data objects, such as Web mining, search marketing, bioinformatics, citation analysis, and epidemiology. A probabilistic model is presented for relational clustering, which also provides a principal framework to unify various important clustering tasks including traditional attributes-based clustering, semi-supervised clustering, co-clustering and graph clustering. The model seeks to identify cluster structures for each type of data objects and interaction patterns between different types of objects. Under this model, parametric hard and soft relational clustering algorithms are provided under a large number of exponential family distributions. The algorithms are applicable to relational data of various structures and at the same time unify a number of state-of-the-art clustering algorithms: co-clustering algorithms, the k-partite graph clustering, and semi-supervised clustering based on hidden Markov random fields.

System and method for probabilistic relational clustering

In a corpus of scientific articles such as a digital library, documents are connected by citations and one document plays two different roles in the corpus: document itself and a citation of other documents. A Bernoulli Process Topic (BPT) model is provided which models the corpus at two levels: document level and citation level. In the BPT model, each document has two different representations in the latent topic space associated with its roles. Moreover, the multi-level hierarchical structure of the citation network is captured by a generative process involving a Bernoulli process. The distribution parameters of the BPT model are estimated by a variational approximation approach.

Knowledge discovery from citation networks

The present invention features a qualitative method to detect independent motion revealed in successive frames of a compressed surveillance MPEG video stream using linear system consistency analysis without decompression of the stream, identifying the segments containing independent motion in a real-time or faster manner, for the retrieval of these segments. The linear system is constructed using the macroblocks of MPEG compressed video frames. The normal flow value of the macroblock is obtained by taking the dot product between the macroblock gradient vector, computed by averaging the four block gradient vectors, and the motion vector of this macroblock. The normal flow value is filtered for inclusion in the linear system, and the statistic of the matrices of the resulting linear system is determined, filtered to screen out false negatives and outliers, and used to determine the presence or absence of independent motion.

Growing community of inventors

Vestal, NY, United States of America

Zhongfei Zhang