Hokkaido University · 컴퓨터과학
마사후루 요시오카 교수의 연구실은 정보 검색, 자연어 처리 및 지능형 설계 시스템 분야에서 핵심적인 연구를 수행하고 있습니다. 특히 법적 텍스트 간의 추론과 정보 추출(예: COLIEE 대회 기반)을 위한 BERT 기반의 앙상블 및 데이터 증강 기법, 뉴스 기사 요약 생성에서의 이벤트 기반 문장 중요도 평가 및 중복 제거 기법 등 자연어 이해 및 처리 기술에 초점을 맞추고 있습니다. 또한, 사용자 정보 필요를 정확히 반영할 수 있는 불리안 쿼리 최적화 및 지능형 쿼리 재작성 기법을 통해 검색 시스템의 정확성과 포괄성을 높이는 데 기여하고 있습니다. 설계 지식의 체계적 정형화 및 컴퓨팅 가능한 모델링을 통한 지능형 CAD 환경 개발도 중요한 연구 주제입니다.
표시된 성과는 수집된 데이터 기준으로 산출되며, 일부 차이가 있을 수 있습니다.
The Competition on Legal Information Extraction/Entailment (COLIEE) statute law legal textual entailment task (task 4) is a task to make a system judge whether a given question statement is true or not by provided articles. In the last COLIEE 2020, the best performance system used bidirectional encoder representations from transformers (BERT), a deep-learning-based natural language processing tool for handling word semantics by considering their context. However, there are problems related to th
In this paper, we describe our method to generate summary from multiple news articles. Since most of news articles report several events and these events are refereed with following articles, we use this event reference information to calculate importance of a sentence in multiple news articles. We also propose a method to delete redundant description by using similarity of events. Finally we discuss its effectiveness based on the evaluation result.
Even though a Boolean query can express the information need precisely enough to select relevant documents, it is not easy to construct an appropriate Boolean query that covers all relevant documents. To utilize a Boolean query effectively, a mechanism to retrieve as many as possible relevant documents is therefore required. In accordance with this requirement, we propose a method for modifying a given Boolean query by using information from a relevant document set. The retrieval results, howeve
Abstract An intelligent CAD system is not merely a set of intelligent design tools, but rather it must be an intelligent, integrated design environment. This requests it must be equipped with a large scale knowledge base in which design knowledge is intensively and systematically stored. To do so, design knowledge must be systematically formalized, made computable, and organized. The present paper investigates fundamental issues of systematization of design knowledge. Design knowledge has two ca
When implementing an IR system that can support comprehensive searches from a wide variety of documents, it is crucial to have a mechanism for selecting appropriate query formulations. For this purpose, many IR systems can modify query terms by estimating the user's information need. However, because modified query terms are usually represented in a complicated form, it is difficult to judge how appropriate they are. In this research, we assume all relevant documents should contain words that co
Even though a Boolean query can express the information need precisely enough to select relevant documents, it is not easy to construct an appropriate Boolean query that covers all relevant documents. To utilize a Boolean query effectively, a mechanism to retrieve as many as possible relevant documents is therefore required. In accordance with this requirement, we propose a method for modifying a given Boolean query by using information from a relevant document set. The retrieval results, howeve
Abstract Most of the previous research efforts for design process modeling had such assumptions as “design as problem solving,” “design as decision making,” and “design by analysis,” and did not explicitly address “design as synthesis.” These views lack notion and understanding about synthesis. Compared with analysis, synthesis is less understood and clarified. This paper discusses our fundamental view on synthesis and approach toward a reasoning framework of design as synthesis. To do so, we ob
