RESEARCH
Current Research
LEARNING DRIVER'S PATTERNS AND BEHAVIOR WITH MACHINE LEARNING
This research looks into big data mining and analysis of BMW car data obtained from BMW Connected, using machine learning algorithms for learning and predicting destinations. Also this looks into analytics for usage of machine learning features and user engagement, personalized individual preferences (like for example seat heating), adaptive location data collection, improving POI accuracy, identifying mobility patterns through activity data, crowdsourcing parked locations, deep learning for vehicle diagnostics, and improved fuel prediction.
Publications:
Peter Wolf, Alvin Chin and Bernard Baker. Unsupervised Data-driven Automotive Diagnostics with Improved Deep Temporal Clustering, In Proc 90th IEEE Vehicular Technology Conference 2019-Fall (VTC 2019-Fall), 22-25 Sept. 2019, Hawaii, USA.
Alvin Chin, Peter Wolf and Jilei Tian. A Cloud IoT Edge Framework for Efficient Data-Driven Automotive Diagnostics, In Proc 90th IEEE Vehicular Technology Conference 2019-Fall (VTC 2019-Fall), 22-25 Sept. 2019, Hawaii, USA.
Olav Laudy, Johann Prenninger, Alvin Chin and Jilei Tian. Toward Building an Individual Preference Model for Personalizing Settings in the Vehicle, In Proc 88th IEEE Vehicular Technology Conference 2018- Fall (VTC 2018-Fall), 27-30 Aug. 2018, Chicago, USA. (pdf)
Previous Research
USER BEHAVIOUR AND RECOMMENDATIONS OF WEB CONTENT
This research looks into analyzing the user behaviour of users browsing the web using Nokia's Xpress Internet browser using big data analytics, as well as creating recommendations of web content based on user's browsing history, trending topics, social interactions and serendipity. Nokia released Nokia Xpress Now, a web app that helps users discover trending and personalized content on the web and uses our recommendation algorithms.
Publication:
Ming He, Alvin Chin, Enhong Chen, and Jilei Tian, “Efficient personalized recommendation of web content using an EM-based clustering method”, Accepted to the 14th International IEEE Conference on Computer and Information Technology (CIT 2014), 8 pages, 11-13 Sept. 2014, Xi'an, China, IEEE Computer Society (pdf)
EVENT-BASED SOCIAL NETWORKING
This research looks into analyzing the user behaviour of users using an event-based system in Douban, a Chinese website for events and how offline influence online. We look into how event properties can affect the attendance of users to an event and a user's social networking behaviour in particular following behaviour. This is joint work with Cornell University.
Publication:
Bin Xu, Alvin Chin and Dan Cosley. "On How Event Size and Interactivity Affect Social Networks". In CHI '13 Extended Abstracts on Human Factors in Computing Systems (CHI EA '13). ACM, New York, NY, USA, 865-870. (pdf)
EPHEMERAL SOCIAL NETWORKING AND MOBILE SOCIAL NETWORKING WITH NOKIA FIND & CONNECT
Nokia Find & Connect is a research project that was created in the Growth Economies Lab Beijing of Nokia Research Center that allows people to find others easily in a conference or workplace environment and then connect with those people from offline to online using physical proximity encounters and homophily. We created a mobile client and system for conferences that supported any mobile device and deployed this at UbiComp 2011 and UIC 2010. We also did a pilot at SXSW 2011 for Nokia bloggers with press coverage here and from SymbianTweet. For the workplace environment, we deployed this in the Nokia office in Beijing and conducted trials in White Plains, NY and Helsinki, Finland. With Nokia Find & Connect, the idea is to automatically collect and identify ad-hoc opportunistic social networks offline during the conference or workplace, that last for a temporary period of time, at a specific point in time at a location which we call ephemeral social networks.
Some of the research topics that were investigated include:
- Identifying and creating ephemeral social networks
- User behavior in offline and online social networks and transition from online to offline
- Inferring social community intelligence from ephemeral social networks
- Friend recommendation using physical context and content
For more information, please read our papers in the Publications section, check out my slides on a talk I did on mobile social networking and our forthcoming book chapter.
ACTIVITY-BASED SOCIAL NETWORKING WITH LEXIANG
LeXiang is a project that was created in the Growth Economies Lab Beijing of Nokia Research Center that allows people to create offline activities for recording photos during the activity and recording the participants that attended the activity. It enables social networking before, during and after an activity. People can find activities that are near them, find the people that attended the activity and build ephemeral social networks. The concept of LeXiang was presented in the YOCSEF forum in Beijing in December 2011.
Publications:
Alvin Chin, Wei Wang, and Xia Wang, "LeXiang: Share the Joy of Reliving Activities from Offline to Online", Accepted to the 7th International IEEE Conference on Cyber, Physical and Social Computing (CPSCom 2014), 8 pages, 1-3 Sept. 2014, Taipei, Taiwan, IEEE Computer Society.
Alvin Chin, Wei Wang, and Xia Wang. “LeXiang: Record, Share and Connect to Relive Activities from Offline to Online”. In Proc. of the Second International Symposium of Chinese CHI (Chinese CHI '14). ACM, New York, NY, USA, 2 pages (Poster)