人工知能学会論文誌
検索
OR
閲覧
検索
29 巻 , 3 号
選択された号の論文の6件中1~6を表示しています
    • |<
    • <
    • 1
    • >
    • >|
一般論文
原著論文
  • 谷藤 拓麻, 松井 俊浩, 松尾 啓志
    29 巻 (2014) 3 号 p. 277-287
    公開日: 2014/04/01
    ジャーナル フリー
    Distributed Constraint Optimization Problem (DCOP) is a basic framework of cooperative problem solving in multi-agent systems. A number of distributed resource allocation problems including sensor networks, smart grids and disaster response tasks are formulated as DCOPs. Since DCOPs are generally NP-hard, incomplete algorithms are practical for large scale applications. DALO is an incomplete algorithm that guarantees the solution quality based on k/t-optimality. The k-optimality defines local optimality criterion based on the size of the group of deviating agents. On the other hand, the t-optimality is based on a group of surrounding agents within a fixed distance of a central agent. In the recent study, C-optimality has been introduced to generalize those criteria. The C-optimality defines criteria for local optimality in any arbitrary regions. As another type of optimality criteria, the p-optimality that is based on the induced width of pseudo-trees on constraint networks has been proposed. With p-optimality, the original problem is approximated by removing back edges of the pseudo-tree. Both types of incomplete algorithms have different week points. Since DALO is based on local optimality criteria, its solution quality depends on limited information (e.g. agent’s values and constraints) within local regions. The solution quality of the p-optimal algorithms decreases when constraint graph consists of many cycles to be removed. In this paper, in order to achieve both lower computational complexity and better solution quality, we propose an integrated solution method based on both types of optimality criteria. Namely, we use information of the incomplete algorithm based on p-optimality and besides another method of C-optimality. Hence our aim is to employ complementary effects of both incomplete algorithms. Empirical results show that our integrated solution method obtain better solution quality than existing incomplete algorithms.
    抄録全体を表示
  • 大島 直樹, 山口 雄大, デシルバ ラビンドラ, 岡田 美智男
    29 巻 (2014) 3 号 p. 288-300
    公開日: 2014/04/04
    ジャーナル フリー
    Understanding why people treat simple geometric animations like real agent which has intention to interact with people even if its geometry is artificial thing will aid the `agency' problems of human-agent interaction. This paper explores effects of treating simple geometric animations as a real participant to facilitate multi-party conversation in social interaction. Observational study was conducted with groups of two or three persons using simple circle (sociable spotlight) which moves based on dynamic information in the current multi-party conversation, with the goal of discovering how participants are utilizing the behaviors of sociable spotlight as an other party for organizing the conversational sequences in talk-in-interaction. In addition, we motivated to explore how the sociable spotlight is embedded within the organization of conversation and how the user's behaviors are changed according to the sociable spotlight's behaviors by investigate through conversation analysis of a video-recording. Finally, we conclude how the agency of artificial things constructed in multi-party conversation from minimal designing point of view.
    抄録全体を表示
  • 小島 諒介, 佐藤 泰介
    29 巻 (2014) 3 号 p. 301-310
    公開日: 2014/04/04
    ジャーナル フリー
    In this paper we propose a new plan recognition method from observations of incomplete action sequences by regarding them as prefixes in a probabilistic context-free grammar (PCFG). In previous work that uses a PCFG for plan recognition, the PCFG receives a sentence, i.e. an observation of complete action sequences to recognize the plan behind it. However, when we deal with real plan recognition problems such as the Web access log analysis, we often cannot have complete sequences of actions and the traditional PCFG approach is not applicable. To overcome this difficulty, we extend the probability computation of PCFGs to prefix probability computation though it requires an infinite sum of probabilities. We applied the proposed method to infer the intended goals of Web site visitors from the online and partial observations of their actions. Also we compared the performance of plan recognition from observations of initial sequences of visitors' actions with that from full observations.
    抄録全体を表示
  • 桝屋 啓志, 溝口 理一郎
    29 巻 (2014) 3 号 p. 311-327
    公開日: 2014/06/05
    ジャーナル フリー
    Gathering information using a computer has become more and more important as a component of research studies in life science. Recently, ontology provides us with one of the most important means of processing varieties of data and representing knowledge models. Currently, several biomedical ontologies have been constructed with the aim of integrating a variety of information produced by different fields of biology. Therefore, development of a common model/ontology for genes is one of the key issues in bioinformatics studies. However, an ontology fully representing multiple aspects of a gene is still not available. In this study, we dissected the biological roles of a gene and built an ontology that represents a consistent data model of the basic concepts of genetics, including genes, alleles, nucleic acid molecules, locus, genotype and phenotype. This is the first ontology to provide a foundation for the construction of a semantic data model for the concept of gene applicable to broad fields of life science including genetics, molecular biology, and population genetics.
    抄録全体を表示
  • 石野 亜耶, 藤井 一輝, 藤原 泰士, 前田 剛, 難波 英嗣, 竹澤 寿幸
    29 巻 (2014) 3 号 p. 328-342
    公開日: 2014/06/05
    ジャーナル フリー
    Travelers planning to visit a particular tourist spot need information about their destination and they often use travel guidebooks (guidebooks) to collect this information. However, guidebooks lack specific information, such as first-hand accounts by users who have visited the specific destination. To compensate for the lack of such information, we focused on travel blog entries (blog entries) and archives of answered question (QA archives). In this paper, we propose a method for enriching guidebooks by aligning with blog entries and question answering archives. This is a three-step method. In Step1, we classify pages of guidebooks, blog entries and QA archives into five types of content, such as "watch" and "eat." In Step 2, we align each blog entry and QA archive with guidebooks by taking these content types into account. In Step 3, we align each blog entry and QA archive with individual pages in guidebooks. To investigate the effectiveness of our method, we conducted a few experiments. Accordingly, 82.0% of blog entries and 77.0% of QA archives were judged to be helpful for travelers. Finally, we constructed a prototype system that provides enriched guidebooks.
    抄録全体を表示
  • 長井 拓馬, 兼岩 憲
    29 巻 (2014) 3 号 p. 343-355
    公開日: 2014/06/05
    ジャーナル フリー
    In this paper, we propose an algorithm for ALCH(D) concept learning from RDF data using minimal model reasoning. This algorithm generates concept expressions in the Description Logic ALCH(D) by giving background knowledge and positive and negative examples in the RDF form. Our method can be widely applied to RDF data on the Web, as background knowledge. An advantage of the method for RDF data is that reasoning on RDF graphs is tractable compared to logical reasoning for OWL data. We solve the problem that RDF data cannot be directly applied to the concept learning due to its less expressive power, speci.cally, the lack of negative expressions. In order to construct expressive ALCH(D) concepts from less expressive RDF data in the concept learning, we introduce (nonmonotonic) inference rules based on minimal model reasoning which derive implicit subclass and subproperty relations from the background knowledge in the RDF form. We prove the soundness, completeness and decidability of the nonmonotonic RDF reasoning in the minimal Herbrand models for RDF graphs. The process of concept learning is divided in two parts: (i) concept generation and (ii) concept evaluation. In the concept generation, minimal model reasoning enables us to derive complex concepts consisting of negation, conjunction, disjunction and quanti.ers and to exclude inconsistent concepts. In the concept evaluation, we evaluate hypothesis concepts with class and property hierarchies where minimal model reasoning is used for expressing more speci.c concepts as the answer for learning. We implement a system that learns some ALCH(D) concepts describing the features of given examples.
    抄録全体を表示
    • |<
    • <
    • 1
    • >
    • >|
feedback
Top