Shunsuke Kozawa

Advice Extraction from Web for Providing Prior Information Concerning Outdoor Activities

Conventional context-aware recommendation systems do not provide information before user action, although they provide information considering users’ ongoing activity. However, users want to know prior information such as how to go to their destination or get necessary items when they plan to do outdoor activities such as climbing and sightseeing. It takes time to collect the prior information since it is not so easy to appropriately find them. This paper proposes a method for extracting prior advices from the web. The method first identifies whether a given sentence is an advice or not. Then the method identifies whether the sentence is a prior advice or not if the sentence is identified as advice. In this paper, we will show availability of the proposed method through our experimentation. We also developed a system for providing prior information using the proposed method.

Revised catalogue specifications of speech corpora with user-friendly visualization and search system

It is well known that speech corpora are indispensable to speech research; several data centers of speech corpora have been set up worldwide in order to meet this demand that serve as a repository for various speech corpora. However, they use different specification systems for their corpora, and so it is difficult for speech corpora users to compare and select suitable corpora. It would be more convenient for the users if each data center used a common specification system for describing its corpora. Based on this idea, we have already proposed a set of specification attributes and items as the first step towards standardization, but the scale of the retrieval system was limited. This paper introduces a revised version of the speech corpora specification attributes and items to be connected with the large-scale metadata database “SHACHI” combined with the “Concentric Ring View (CRV) System” to improve the user interface.

Automatic Collection of Useful Phrases for English Academic Writing

English academic writing is indispensable for researchers to present their own research achievement. It is hard for non-native researchers to write research papers in English. They often refer to phrase dictionaries for academic writing to know useful expressions in academic writing. However, lexica available in the market do not have enough expressions and example sentences to serve the purpose since the lexica are created by hand. In order to respond to the demand for the better lexica, this paper proposes a method for extracting useful expressions automatically from English research papers. The expressions are extracted from research papers based on four characteristics of the expressions. The extracted expressions are classified into five classes; “introduction”, “related work”, “proposed method”, “experiment”, and “conclusion”. In our experiment using 1,232 research papers, our proposed method achieved 57.5% in precision and 51.9% in recall. The f-measure was higher than those of the baselines, and therefore, we confirmed the validity of our method. We developed a phrase search system using extracted phrasal expressions to support English academic writing.

Design and collection of ontological metadata for enhancing interoperability of language resources

This paper describes the design and implementation of a large scale ontological database named SHACHI, storing detailed metadata on language resources (LRs) in Asian and Western countries. SHACHI has been constructed to enhance the interoperability of LRs, that is, to effectively combine LRs, to systematically store LR metadata, to provide a common infrastructure for web services, to investigate languages, tag sets, and formats compiled in LRs, and to ultimately utilise all these factors for more efficient development of LRs. This ontological metadata database, containing more than 2,000 compiled LRs such as corpora, dictionaries, thesauruses and lexicons, has an aspect of an archive of a large scale metadata of LRs, and its website is now open to the public and accessible to all internet users. SHACHI metadata set is an extended version of OLAC metadata set which conforms to Dublin Core metadata element set. This paper first presents the methodologies to systematically store LR metadata and efficiently LR catalogues, and then explains the structure of the ontological metadata database, as well as the realisation of the LR catalogue search tool. The usefulness of the ontology search function has been investigated.

Acquisition of know-how information from web

A variety of know-how such as recipes and solutions for troubles have been stored on the Web. However, it is not so easy to appropriately find certain know-how information. If know-how could be appropriately detected, it would be much easier for us to know how to tackle unforeseen situations such as accidents and disasters. This paper proposes a promising method for acquiring know-how information from the Web. First, we extract passages containing at least one target object and then extract candidates for know-how from them. Then, passages containing the know-how are discriminated from non-know-how information considering each object and its typical usage.

Automatic Extraction of Phrasal Expressions for Supporting English Academic Writing

English academic writing is not easy for non-native researchers. They often refer to lexica of phrases on English research papers to know useful expressions in academic writing. However, lexica on sales do not have enough amount of expressions. Therefore, we propose a method for automatically extracting useful expressions from English research papers. We found four characteristics of the expressions by analyzing the existing lexicon of phrases on English research papers. The expressions are extracted from research papers based on statistical and syntactic information. In our experiment using 1,232 research papers, our proposed method achieved 57.5% in precision and 51.9% in recall. The f-measure was higher than the baselines, and therefore, we confirmed the feasibility of our method.