I am now a "visiting engineer" with Pan Zhang at the Institute of Theoretical Physics, Chinese Academy of Sciences in Beijing. I hold a B.S. in Life Science in National Taiwan University, Taiwan. My research interests include statistical learning methods in network models, computational social science, and web technologies.
I was a full-stack engineer at Sensoro Technology in Beijing from 2015 to 2016. I also worked with Tang Jie at Tsinghua University, Cheng-Te Li at National Cheng Kung University, Chia-Lung Hsieh at Academia Sinica, and with Yuan-Chung Cheng at National Taiwan University.
- [Jun. 2017] I will attend the NetSci2017 in Indianapolis.
- I will give a talk at the Statistical Inference of Network models (SINM) satellite symposium, as well as contribute a poster in the main conference. Both titled: Community Number Selection in Bipartite Networks.
- [May. 2017] I will attend the Sunbelt Conference in Beijing.
- Accepted poster title: Understanding Online Survey Data as a Bipartite Network: A Case Study on pol.is.
- [Jan. 2017] Moved to work with Pan Zhang in Beijing.
- [Dec. 2016] I am looking for Ph.D. opportunities. Up for any interesting projects. Please contact me.
- [Apr. 2016] My work back in Sensoro is accepted as a poster contribution at the NetSci2016. Thanks Sensoro for generously providing me their data. Looking forward to the conference in Seoul, South Korea.
- [Apr. 2016] Back in academia! I left Sensoro to start my new adventure at Tang's Group, best known for their academic social network search system Aminer. I plan to learn a lot in probabilistic graphical models, data mining methodologies, and social network theories.
- [Oct. 2015] Our work on reasoning noise pollution by mining multimodal geo-social big data in New York City has been published and selected as the Grand Challenge Finalist in ACM-MM’15 in Brisbane, Australia.
- [Aug. 2015] Our work on an analysis and visualization framework for urban construction data had been published in UrbComp’15 in Sydney, Australia.
- [May 2015] Our team, Geomonsters, won the Second Prize at the Taipei City Open Data Hackathon in May.
- [Mar. 2015] I started to work as a web engineer at Sensoro Technology.
Machine Learning with Social Impact
I proposed an analytic framework for data with schema – geo-location, time, and type – simultaneously. We built a demo system using urban road construction open data from Taipei City Government. The system aims to visualize, cluster, and predict road construction dynamics (Yen et al., UrbComp'15). In addition, with my collaborators, I studied urban noise composition inference using geo-social multimedia data, such as Twitter, Foursquare, Flickr, and Gowalla. The system – New York City Urban Noise Diagnotor – was developed (Hsieh et al., MM'15) for the citizens to query the noise distribution of any place in NYC.
I am following the projects in the Data Science for Social Good Fellowship in the University of Chicago, and look forward to initiate some projects of our own in my home country, Taiwan.
Web System for Information Visualization
A good visualization with an analytic core is worth a thousand words. I built (1) a system to identify key influentials in a customer network. It can automatically collect, sort, rank, and visualize WeChat user sharing log data as dynamic social networks (DEMO video). And also (2) a multifaceted selection-and-query system for a large pool of labelled customers (DEMO video). The two works grew into core features of Sensoro's Social CRM system.
Collaborated with Cheng-Te Li, we gave birth to the traditional Mandarin version of "Network Literacy: Essential Concepts and Core Ideas" with the Network Science in Education (NetSciEd) Initiative. We included figures from various research papers, and wrote informative captions in hope to help the reader better understand the topic. We are indebted to the authors that provided us their great work. We hope the book could bring Network Science to the Chinese community (link to project page [Mandarin]).
Network and Dynamics in Biological Systems
[Old works] I studied energy transfer dynamics in light-harvesting networks. We proposed clustered chlorophyll geometries may enhance energy transfer due to quantum coherence. We justified our work in the natural FMO complex (Ai et al., JPCL'13). I also studied single-molecule diffusion in supported lipid membranes using ultrafast florescence microscopy. Specifically, a monomeric avidin-like protein diffusion was filmed and compared to pure lipid diffusion (Jeong et al., Angew Chem'16). In addition, diffusion in a raft-containing reconstituted membranes (Wu et al., Sci Rep'16) is studied. In many cases, sub-diffusion in mini-to-microsecond timescale was observed. These studies help understand single-molecule diffusion in complex environments.
In summary, I am interested in using network methodologies or statistical methods to contribute to the set of algorithms in nature.