Kanji Uchino

M-CAFE: Managing MOOC Student Feedback with Collaborative Filtering

Ongoing student feedback on course content and assignments can be valuable for MOOC instructors in the absence of face-to-face-interaction. To collect ongoing feedback and scalably identify valuable suggestions, we built the MOOC Collaborative Assessment and Feedback Engine (M-CAFE). This mobile platform allows MOOC students to numerically assess the course, their own performance, and provide textual suggestions about how the course could be improved on a weekly basis. M-CAFE allows students to visualize how they compare with their peers and read and evaluate what others have suggested, providing peer-to-peer collaborative filtering. We evaluate M-CAFE based on data from two EdX MOOCs.

Topic-Specific Recommendation for Open Education Resources

Most of the Open Educational Resources are scattered around the Web, and not well described and structured so causing huge problems in their use, search, organization and management. We present a system which can collect and analyze domain-specific contents from massive online learning materials and publications, and automatically identify domain-specific knowledge terms and aggregate them into well-organized topics. The system helps learners who want to learn cutting-edge technology to effectively and efficiently locate appropriate online courses and learning materials using topic-specific recommendation. We further examines the system with real-world data and applications.

Domain-Specific Recommendation by Matching Real Authors to Social Media Users

It is important to discover informative users disseminating fresh and high-quality domain-specific contents over social media in order to keep up-to-date with and learn cutting-edge knowledge, but that is not easy, especially for new learners due to information abundance or even overload. We propose an efficient approach to discover potential informative users by matching real-world authors extracted from the latest domain-specific publications to corresponding social media user accounts. Mutually reinforcing methods are further applied to identify informative users and recommend domain-specific contents in social media. Our experiments on real data from arxiv and twitter are used to verify feasibility and effectiveness of the proposed methods.

Guided Play: Automatic Stereotypical Behavior Analysis and Intervention during Play

Restricted and repetitive behaviors (RRB) are a core symptom and an early marker of Autism Spectrum Disorder (ASD). Despite technologies for detecting certain forms of RRB, assessment and intervention for RRB still heavily rely on professional experience and effort. This paper presents an ongoing investigation of a technology that uses instrumented games or toys as platforms to assess RRB and facilitate behavior intervention during play. The design and implementation of a prototype for the iPad are discussed. The same technology can be applied to tangible objects such as smart toys for a natural player-computer interface.

Mining Domain-Specific Accounts for Scientific Contents from Social Media

This paper proposes a machine learning based approach to automatically create an initial set of domain-specific accounts by matching real-world authors of the latest domain-specific publications to corresponding social media accounts. An efficient approach based on social network analysis is further applied to extend the initial set by finding more domain-specific accounts of various types and filtering out irrelevant general or non-domain-specific accounts. Our experiments on Twitter are used to verify feasibility and effectiveness of the proposed methods.

An Assistant Interface for Finding Query-Related Proper Nouns

Searching information on the Web is now one of the important acts in dairy work because we can obtain several types of information from the Web. For example, we use web search engines to obtain information related to proper nouns, such as famous persons, shops, restaurants, sightseeing places and so on. However, we often obtain irrelevant information with what we want to obtain because most web search engines just return web pages including a posted query. This paper describes a web search engine interface for users finding proper nouns and their relevant information. Our interface selects proper nouns related to a query from a search result with Natural Language Processing technologies. Our interface provides users the following information in addition to original search result: 1) Lists of proper nouns related to a posted query. 2) Links to each proper noun to results includi ng the proper noun. 3) Links for searching new information by using proper nouns as a new query. The experimental results with four subjects have shown that our interface contributes to find target proper nouns.

RSS feed generation from legacy HTML pages

Although RSS demonstrates a promising solution to track and personalize the flow of new Web information, many of the current Web sites are not yet enabled with RSS feeds. The availability of convenient approaches to “RSSify” existing suitable Web contents has become a stringent necessity. This paper presents a system that translates semi-structured HTML pages to structured RSS feeds, which proposes different approaches based on various features of HTML pages. For the information items with release time, the system provides an automatic approach based on time pattern discovery. Another automatic approach based on repeated tag pattern mining is applied to convert the regular pages without the time pattern. A semi-automatic approach based on labelling is available to process the irregular pages or specific sections in Web pages according to the user’s requirements. Experimental results and practical applications prove that our system is efficient and effective in facilitating the RSS feed generation.