Probabilistic latent semantic analysis

About: Probabilistic latent semantic analysis is a research topic. Over the lifetime, 2884 publications have been published within this topic receiving 198341 citations. The topic is also known as: PLSA.

Factorization of Latent Variables in Distributional Semantic Models

Abstract: This paper discusses the use of factorization techniques in distributional semantic models. We focus on a method for redistributing the weight of latent variables, which has previously been shown to improve the performance of distributional semantic models. However, this result has not been replicated and remains poorly understood. We refine the method, and provide additional theoretical justification, as well as empirical results that demonstrate the viability of the proposed approach.

A semantic representation for image retrieval

Abstract: Robust semantic labeling of image regions is a basic problem in representing and retrieving image/video content. We propose an SVM-MRF framework to model features and their spatial distributions, leading towards a "semantic" representation. Eigenfeatures of Gabor wavelet features and Gaussian mixture model are used for feature clustering. Since similar feature vectors in one cluster can come from several different semantic classes, SVM is applied to represent conditioned feature vector distributions within each cluster, and a Markov random field is used to model the spatial distributions of the semantic labels. A semantic layout representation is proposed to describe the semantics of the images. Experiments show that this method can improve semantic labeling and is useful in similarity search.

The Metric of the Latent Variables in a LISREL-IV Analysis

Abstract: A potential source of confusion in the interpretation of a LISREL-IV analysis is the metric of the latent variables. This paper demonstrates that fixing the pattern coefficient of one of the indicators of each latent variable to 1.0 results in an arbitrary and meaningless metric which is usually different for each latent variable. Since many of the parameter estimates such as the pattern coefficients, the factor loadings, and the variance-covariance matrix of the latent variables are a function of the metric of the latent variables, much of the LISREL-IV output may be uninterpretable. This paper demonstrates how to specify a priori the variances of the latent variables in a LISREL-IV analysis.

Geo-location inference on news articles via multimodal pLSA

Abstract: The fast evolution and adoption of social media creates an increasingly need for location based services. Location inference on news or events becomes an essential problem. This paper addresses the problem by extracting location involved topics (geo-topic) using both text content and visual content. This paper proposes a geo-topic extraction framework for geo-location inference, including location name entity recognition, location related image association and a multimodal location dependent pLSA geo-topic model. Experiments have shown that our fused model improves the f-score in geo-location inference by 10% compared with single modality based models.

Quantitative Semantic Analysis and Comprehension by Cognitive Machine Learning

Abstract: Knowledge learning is the sixth and the most fundamental category of machine learning mimicking the brain. It is recognized that the semantic space of machine knowledge is a hierarchical concept network HCN, which can be rigorously represented by formal concepts in concept algebra and semantic algebra. This paper presents theories and algorithms of hierarchical concept classification by quantitative semantic analysis based on machine learning. Semantic equivalence between formal concepts is rigorously measured by an Algorithm of Concept Equivalence Analysis ACEA. The semantic hierarchy among formal concepts is quantitatively determined by an Algorithm of Relational Semantic Classification ARSC. Experiments applying Algorithms ACEA and ARSC on a set of formal concepts have been successfully conducted, which demonstrate a deep machine understanding of formal concepts and quantitative relations in the hierarchical semantic space by machine learning beyond human empirical perspectives.