Tomonari Masada

Enumeration of regular triangulations

Regular triangulations form a meaningful wide subclass of triangulations of points in general dimensions. They can be defined as a natural extension of the Delaunay triangulation and also of lexicographic triangulations, a subclass of triangulations well-known in the theory of oriented mat roids. Moreover, regular triangulations have interesting algebraic aspects in connection with a famous paradigm of computer algebra, Grobner bases. This paper proposes an output-size sensitive and work-space efficient algorithm to enumerate all regular triangulations by reverse search. The algorithm makes full use of the existing results on the secondary polytope [BFS 90, GKZ 94] whose vertices correspond to regular triangulations. These known results are summarized with only using the so-called volume vector, and the algorithm is described in a simple way. Some regular triangulations may not use a point inside the convex hull, which may not be preferable for three-dimensional applications in computer graphics and finite element met hod. Triangulations using all the points are called spanning, and an algorithm is given to enumerate all spanning regular triangulations. The diameter of the secondary pol yt ope is investigated. Preliminary comput ational results are also shown. From the viewpoint of computational geometry, these generalizes the results for planar triangulations to higher-dimensional cases by restricting triangulations to be regular.

Using a knowledge base to disambiguate personal name in web search results

Results of queries by personal names often contain documents related to several people because of the namesake problem. In order to differentiate documents related to different people, an effective method is needed to measure document similarities and to find documents related to the same person. Some previous researchers have used the vector space model or have tried to extract common named entities for measuring similarities. We propose a new method that uses Web directories as a knowledge base to find shared contexts in document pairs and uses the measurement of shared contexts to determine similarities between document pairs. Experimental results show that our proposed method outperforms the vector space model method and the named entity recognition method.

Accelerating Collapsed Variational Bayesian Inference for Latent Dirichlet Allocation with Nvidia CUDA Compatible Devices

Next-Generation Applied Intelligence: 22nd International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, IEA/AIE 2009, Tainan, Taiwan, June 24-27, 2009.

Dynamic hyperparameter optimization for bayesian topical trend analysis

This paper presents a new Bayesian topical trend analysis. We regard the parameters of topic Dirichlet priors in latent Dirichlet allocation as a function of document timestamps and optimize the parameters by a gradient-based algorithm. Since our method gives similar hyperparameters to the documents having similar timestamps, topic assignment in collapsed Gibbs sampling is affected by timestamp similarities. We compute TFIDF-based document similarities by using a result of collapsed Gibbs sampling and evaluate our proposal by link detection task of Topic Detection and Tracking.

Disambiguation of People in Web Search Using a Knowledge Base

Results of queries by personal names often contain documents related to several people because of the namesake problem. In order to differentiate documents related to different people, an effective method is needed to measure document similarities and to find documents related to the same person. Some previous researchers have used the vector space model or have tried to extract common named entities for measuring similarities. We propose a new method that uses Web directories as a knowledge base to find shared contexts in document pairs and uses the measurement of shared contexts to determine similarities between document pairs. Experimental results show that our proposed method outperforms the vector space model method and the named entity recognition method.

Bayesian Multi-topic Microarray Analysis with Hyperparameter Reestimation

This paper provides a new method for multi-topic Bayesian analysis for microarray data. Our method achieves a further maximization of lower bounds in a marginalized variational Bayesian inference (MVB) for Latent Process Decomposition (LPD), which is an effective probabilistic model for microarray data. In our method, hyperparameters in LPD are updated by empirical Bayes point estimation. The experiments based on microarray data of realistically large size show efficiency of our hyperparameter reestimation technique.

Comparing LDA with pLSI as a dimensionality reduction method in document clustering

In this paper, we compare latent Dirichlet allocation (LDA) with probabilistic latent semantic indexing (pLSI) as a dimensionality reduction method and investigate their effectiveness in document clustering by using real-world document sets. For clustering of documents, we use a method based on multinomial mixture, which is known as an efficient framework for text mining. Clustering results are evaluated by F-measure, i.e., harmonic mean of precision and recall. We use Japanese and Korean Web articles for evaluation and regard the category assigned to each Web article as the ground truth for the evaluation of clustering results. Our experiment shows that the dimensionality reduction via LDA and pLSI results in document clusters of almost the same quality as those obtained by using original feature vectors. Therefore, we can reduce the vector dimension without degrading cluster quality. Further, both LDA and pLSI are more effective than random projection, the baseline method in our experiment. However, our experiment provides no meaningful difference between LDA and pLSI. This result suggests that LDA does not replace pLSI at least for dimensionality reduction in document clustering.

Web Page Grouping Based on Parameterized Connectivity

We propose a novel method for Web page grouping based only on hyperlink information. Because of the explosive growth of the Web, page grouping is expected to provide a general grasp of the Web for effective Web search and netsurfing. The Web can be regarded as a gigantic digraph where pages are vertices and links are arcs. Our method is a generalization of the decomposition into strongly connected components. Each group is constructed as a subset of a strongly connected component. Moreover, group sizes can be controlled by a parameter, called the threshold parameter. We call the resulting groups parameterized connected components. The algorithm is simple and admits parallelization. Notably, we apply Dijkstra’s shortest path algorithm in our method.

Using Web Directories for Similarity Measurement in Personal Name Disambiguation

In this paper, we target on the problem of personal name disambiguation in search results returned by personal name queries. Usually, a personal name refers to several people. Therefore, when a search engine returns a set of documents containing that name, they are often relevant to several individuals with the same namesake. Automatic differentiation of people in the resulting documents may help users to search for the person of interest easier. We propose a method that uses Web directories to improve the similarity measurement in personal name disambiguation. We carried out experiments on real Web documents in which we compared our method with the vector space model method and the named entity recognition method. The results show that our method has advantages over these previous methods.

A package for triangulations

Recently, triangulations of a point set have been intensively studied in computational geometry. Triangulations are very important not only from the viewpoint of their applications to computer graphics, numerical analysis but also from a fact that triangulations are yet another fundamental structure in discrete geometry like hyperplane arrangements. Our group has developed many systems concerning triangulations. Specifically,