This paper proposes to extend must-link constrained K-means clustering by introducing dynamic generation of subordinate clusters. When clustering high-dimensional data there is a case where data which should belong to the same cluster form several distinct groups in a data space. In order to handle such a case without using distance metric learning, the proposed method generates subordinate clusters for each data group, which are merged after finishing K-means clustering. Result of a comparison experiment with a baseline method shows the effectiveness of the proposed method in terms of success rate and NMI (normalized mutual information)

This paper aims to introduce a ranking function to context search engine. Context search engine has been developed for answering trend-related queries. In order to achieve a more efficient search, ranking retrived results, which is one of important function of exsisting Web search engines, is expected to be effective also for the context search engine. This paper discusses several features that could be used for ranking function of context search engines, such as intensity and periodicity of temporal change. The result of ranking retrieved results with the intensity of temporal change is also shown.

The goal of our study is to establish a new research topic named "Comic computing." With the spread of small devices like tablet PC and smart phone, a market for e-books has been growing. In particular, expectation for digital comics is so huge that comics account for the largest portion in the sales amount. Under such circumstances, this paper presents a service concepts that can be realized when the comic contents become computable.

Passive information consumption would an adequate type of information behavior for receiving the content of, for example, a news article. It may however be boring in many cases and even painful in some cases, especially when the information content is delivered by employing speech media. The user of a speech-based information delivery system, for example a text-to-speech system, usually cannot interrupt the ongoing information flow, inhibiting her/him to confirm some part of the content, or to pose an inquiry for further information seeking. We thus argue that spoken dialogue is a suitable media for enabling interactive information access that coordinate passive information consumption and active information seeking. This paper shows that a carefully designed spoken dialog system could remedy these undesirable situations, and further enables an enjoyable conversation with the users. The key technologies to realize such an attractive speech-based interactive information access system are: (1) pre-compilation of a dialog plan based on the analysis of a source content, and (2) the dynamic recognition of user's state of understanding and interests during the course of conversation. This paper illustrates technical views to implement these functionalities, and discusses a dialog example to exemplify our approach.

This paper reports the preliminary study on the development of monitoring support system for stream data. It is supposed that stream data such as online news has to be monitored during break of user's primary job. If a user check a stream data at wrong time, the efficiency of his/her primary job would go down. In order to help a user to monitor stream data, we are developing a system that gives a user a clue for determining the timing to monitor with using a dynamic bar chart. This paper reports the result of preliminary experiment, in which the effect of color of bars and individual differences on the timing decision is investigated.