Katsuhiko Toyama

MAGNET: ad hoc network system based on mobile agents

This paper proposes an ad hoc network system based on mobile agents. We regard some special mobile agent as an implementation of the network protocol. Mobile agents can move over the network while being carried by protocol agents. Implementation of the protocols by the mobile agents enables the dynamic extension of the network. We introduce agent replication for the simple development of the inter-agent communication. Agent hierarchy enables the component software design. We also describe the implementation of the ad hoc network protocol DSR using our prototype system MAGNET that is based on the framework.

PLSI utilization for automatic thesaurus construction

When acquiring synonyms from large corpora, it is important to deal not only with such surface information as the context of the words but also their latent semantics. This paper describes how to utilize a latent semantic model PLSI to acquire synonyms automatically from large corpora. PLSI has been shown to achieve a better performance than conventional methods such as tf·idf and LSI, making it applicable to automatic thesaurus construction. Also, various PLSI techniques have been shown to be effective including: (1) use of Skew Divergence as a distance/similarity measure; (2) removal of words with low frequencies, and (3) multiple executions of PLSI and integration of the results.

Metric Learning for Synonym Acquisition

The distance or similarity metric plays an important role in many natural language processing (NLP) tasks. Previous studies have demonstrated the effectiveness of a number of metrics such as the Jaccard coefficient, especially in synonym acquisition. While the existing metrics perform quite well, to further improve performance, we propose the use of a supervised machine learning algorithm that fine-tunes them. Given the known instances of similar or dissimilar words, we estimated the parameters of the Mahalanobis distance. We compared a number of metrics in our experiments, and the results show that the proposed metric has a higher mean average precision than other metrics.

Bootstrapping-Based Extraction of Dictionary Terms from Unsegmented Legal Text

Recent demands for translating Japanese statutes into foreign languages necessitate the compilation of standard bilingual dictionaries. To support this costly task, we propose a bootstrapping-based lexical knowledge extraction algorithm Monaka , to automatically extract dictionary term candidates from unsegmented Japanese legal text. The algorithm is based on the Tchai algorithm and extracts reliable patterns and instances in an iterative manner, but instead uses character n -grams as contextual patterns, and introduces a special constraint to ensure proper segmentation of the extracted terms. The experimental results show that this algorithm can extract correctly segmented and important dictionary terms with higher accuracy compared to conventional methods.

Sync/Trans: Simultaneous Machine Interpretation between English and Japanese

This paper describes Sync/Trans, an incremental spoken language translation system. The system has been being developed for efficiently translating a spontaneous speech dialogue between an English speaker and a Japanese speaker. Its purpose being to behave as a simultaneous interpreter, the system produces the target output synchronously with the source input. Sync/Trans has the following features: (1) the system consists of modules that work in a synchronous fashion, (2) the system translates the source language possibly word-by-word according to the appearance order, (3) the system utilizes grammatically ill-formed expressions for the speech output, and (4) the system corrects the grammatical ill-formedness of the speech input at a pretty early stage. An experimental system for translating EngUsh speech into Japanese speech has been implemented. A few experimental results have shown Sync/Treins to be a promising system for simultaneous interpretation.

Extraction of Legal Definitions and Their Explanations with Accessible Citations

The aim of this paper is to produce a Japanese legal terminology consisting of legal terms and their explanations that includes accessible citations. Although we have succeeded in finding over 14,000 terms with high precision, 23.1 percent of the correct explanations included citations that were inaccessible due to context-dependent format. We propose a method for revising explanatory sentences that takes into account XML-tag annotation for context-independent format for all citations. The effectiveness of this method is confirmed by our experimental results.

Design and compilation of syntactically tagged corpus of japanese statutory sentences

This paper describes how to analyze Japanese statutory sentences. Although statutory sentences have many technical terms and characteristic structures, the design of the general Japanese corpus has no tags to handle such structures and usual Japanese parsers cannot analyze them correctly. Thus, we propose a new design of syntactic tags for statutory sentences and develop a support tool that corrects the output of a parser. In this paper, we focus on dependency structure of sentences and present an overview of the design and compilation of the corpus of Japanese statutory sentences.

Ad hoc network system based on infrared communication

Recently, it becomes popular to use small size computers such as notebook computers or PDAs in the mobile environment. It sometimes happens that several computers meet at the same place such as meeting rooms or conference sites. In such environment, there are demands to make a direct communication among mobile computers. Route maintenance and host enumeration are key requirement for such an ad hoc network. In this paper, we propose a network system based on infrared communication. Our system solves host enumeration as well as route maintenance using diffusing computation. We describe autonomous communication protocol for the ad hoc network and an implementation of mobile system using the protocol.

Evaluation metrics for consistent translation of japanese legal sentences

We propose new translation evaluation metrics for legal sentences. Since most previous metrics, that have been proposed to evaluate machine translation systems, prepare human reference translations and assume that several correct translations exist for one source sentence. However, readers usually believe that different translations denote different meanings, so that the existence of several translations of one legal expression may confuse them. Therefore, since translation variety is unacceptable and consistency is crucial in legal translation, we propose two metrics to evaluate the consistency of legal translations and illustrate their performances by comparing them with other metrics.

An Experiment on Japanese-Uighur Machine Translation and Its Evaluation

This paper describes an evaluation experiment about a Japanese-Uighur machine translation system which consists of verbal suffix processing, case suffix processing, phonetic change processing, and a Japanese-Uighur dictionary including about 20,000 words. Japanese and Uighur have many syntactical and language structural similarities, including word order, existence and same functions of case suffixes and verbal suffixes, morphological structure, etc. For these reasons, we can consider that we can translate Japanese into Uighur in such a manner as word-by-word aligning after morphological analysis of the input sentences without complicated syntactical analysis. From the point of view of practical usage, we have chosen three articles about environmental issue appeared in Nippon Keizai Shinbun, and conducted a translation experiment on the articles with our MT system, for clarifying our argument. Here, we have counted the correctness of phrases in the Output sentences to be evaluating criteria. As a results of the experiment, 84.8% of precision has been achieved.