コロキアムB発表

日時: 9月24日(木)3限(13:30~15:00)


会場: L2

司会: 藤本 雄一郎
磯田 祥吾 M, 2回目発表 ユビキタスコンピューティングシステム 安本 慶一, 中村 哲, 諏訪 博彦(特任准教授), 松田 裕貴
title: On-site tourism planning methods that enable you to visit the must-see spots at better times
abstract: Most of the existing studies on route recommendation are based on the satisfaction with the next spot only, but they do not take into account the next and subsequent spots, which limits the sightseeing in the following spots. There is a possibility that a trade-off relationship between satisfaction with the next place to visit and the expected satisfaction with the next and subsequent places to visit may occur. The satisfaction level of visitors differs depending on the time of the day they visit the same spot, so visiting at a better time of the day leads to the improvement of overall satisfaction of tourism. Taking these two points into account, it is desirable to provide users with tourist routes, but to the best of our knowledge, no such system exists. In this paper, in order to take the above into account, we first formulate a tour score consisting of three components: the static and dynamic tourist context of the next spot, and the expected satisfaction level from the next and subsequent spots to be visited. The problem of calculating the tour route that maximizes the tour score (the sequence of the next and subsequent spots to be visited) is NP-hard and in order to find a quasi-optimal solution onsite, we developed three algorithms based on the greedy method: (1) a greedy method that considers only the next spot, (2) a greedy method that considers the entire tour time (3) The greedy method is proposed to extend the whole sightseeing time and the scope of exploration. In order to investigate the usefulness of the proposed algorithm, three algorithms were applied to 20 spots to be visited in Higashiyama-ku, Kyoto, and the output solution was found to be superior to that of the Kyoto model route, with a computation time of 1.9(s), 15.9(s), and 7766.6(s), respectively.
language of the presentation: Japanese

 
松井 智一 M, 2回目発表 ユビキタスコンピューティングシステム 安本 慶一, 中村 哲, 諏訪 博彦(特任准教授)
title: Daily Activity Sensing and Activity Recognition with Easy-To-Deploy Living Activity Sensing System
abstract: Emergence of smart appliances and high performance IoT devices is promoting studies on more functional and intelligent home services using these devices. In Japan, we are facing the problem of aging population and declining birthrate, hence it is urgent to develop technologies to improve resident’s QoL and monitor the elderly through home services based on the activity recognition technology. However, an activity recognition system in general requires many types/number of sensors and hence it is difficult to deploy and operate it in general households. In this paper, we propose a system consisting of low-cost and easy-to-deploy sensors that collects data of resident’s activities of daily living (ADL). The system was deployed in actual homes of senior citizens and collected ADL data for two months. We also estimated the ADLs from the collected data by DNN(Deep Neural Network). As a result, ADLs could be estimated at high F measure of 59% and hence we found that the proposed system has high applicability to actual services.
language of the presentation: Japanese
発表題目: 設置が容易なセンシングシステムによる生活行動データ収集と行動認識
発表概要: 家電のスマート化やIoT機器の高性能化を背景に,宅内サービスの高機能化が研究されている.特に,我が国では少子高齢化が進行していることから,生活行動推定技術を用いた宅内サービスによる居住者のQoL向上や,高齢者の見守りが切望されている.一方で,生活行動推定は多種・多数のセンサを要するため,一般家庭への設置・運用が難しいという課題がある.本研究では,安価かつ設置・運用が容易なセンサからなる生活データ収集システムを構築した.構築したシステムを一般の高齢者家庭に設置し,1カ月間の生活データ収集を行った.収集したデータを基にDNN(Deep Neural Network)による分析を行った結果,約59%の精度で生活行動の認識が可能であることを確かめた.
 
RACH NIKLAS D, 中間発表 ユビキタスコンピューティングシステム 安本 慶一, 中村 哲, 諏訪 博彦, 松田 裕貴
Title: Towards Flexible Argumentation with Conversational Agents
Abstract: The ability to argue about topics of interest and to adapt the argumentation to a specific interlocutor are crucial aspects of problem solving between humans. It is therefore desirable to equip conversational agents with these cognitive capacities, especially in view of applications like smart environments and virtual assistants. This talk gives an overview over recent research into both directions. The diversity of topics is addressed by investigating the applicability of argument search engines in dialogue systems as a way of retrieving a vast amount of arguments for various topics. Based on data collected in this first step, an estimation of subjective argument quality aspects from social signals is proposed as an approach to user-adaptive argumentation in the second part.
language of the presentation: English
 
佐々木 皓大 M, 2回目発表 ユビキタスコンピューティングシステム 安本 慶一, 中村 哲,  諏訪 博彦
title: Designing a Support System for Risk Reduction in Stock Investment using Social Media Messages
abstract: In recent years, there has been an increase in the number of people who choose stock investment for asset management. When doing actual investment and asset management, reducing risks is a very important issue. The Nikkei VI is a risk indicator of the Nikkei Stock Index, and predicting the rise of this indicator may help investors to reduce their risk. On the other hand, investors share their opinions and thoughts in social media. Since stock prices are affected by many factors, including macroeconomics and various news, we assume that the topics discussed in social media are important in the prediction of stock movements. In this study, I propose a system to predict the rise of the Nikkei VI using social media topics for supporting investors' risk reduction. Moreover, the expected goals and our current progress will also be reported here.
language of the presentation: Japanese