About
About Xinmin Chu
I work at the intersection of computational science, statistics, and economics. My research asks how defensible structure can be recovered from fragmented, duplicated, multiview, or partially observed data—and how the resulting decisions can remain auditable.
Research path
I am a Ph.D. candidate in Computational Science at the University of Massachusetts Boston, with expected completion in December 2026. I previously earned an M.S. in Statistics from the University of New Mexico and a Ph.D. in Economics from Xi'an Jiaotong University. This path connects measurement, inference under noise, computation, and the practical consequences of analytical choices.
My current work centers on entity resolution and deduplication, multiview representation and fusion, similarity graphs and clustering, threshold selection, and reproducible evaluation. Related work includes tensor methods, missing-data regression, and applied statistical analysis in healthcare research. The public artifact of the entity-resolution program is erbot, an open-source R toolkit; publications are listed on the Research page.
Teaching and learning
I taught economics, e-commerce and the digital economy, econometrics, managerial economics, industrial economics, and international business at Jiangsu Normal University, and later taught introductory statistics and supported R and Python labs at the University of New Mexico. I have also supported instruction in numerical methods, linear algebra, and machine learning at the University of Massachusetts Boston.
My current education interests include AI-supported practice for learners who lack access to expert partners, and privacy-conscious systems that use interaction evidence to help learners and educators understand progress without reducing people to opaque inferred traits.
Building
I build research-informed software to test ideas in usable form. ERBOT is an R-based entity-resolution research program; ChuKits is a bilingual field guide for identifying and understanding technical tools and equipment; DreamReality explores traceable transformations from dream narratives to structured visual media. Each project carries an explicit maturity label so that an idea, prototype, research artifact, and public service are not presented as the same thing.
Earlier applied research and innovation
Before the current computational-science work, I worked on applied engineering problems in agricultural and environmental monitoring in China. That work produced the patent grants below. It predates the research above and is listed separately from it.
便携式手动全方位多角度观测架
Portable manual omnidirectional multi-angle observation standDraft translation
一种基于物联网的养殖场环境监测手持设备
Handheld device for farm environmental monitoring based on the Internet of ThingsDraft translation
一种养殖业实时大数据监测系统
Real-time big-data monitoring system for the livestock and aquaculture industryDraft translation
一种处理集约化养牛场排泄物的真空厌氧处理设备
Vacuum anaerobic treatment equipment for manure from intensive cattle farmsDraft translation
一种厌氧处理集约化养牛场排泄物的方法
Method for anaerobic treatment of manure from intensive cattle farmsDraft translation
Historical patent grants in China. Each entry states the grant as recorded at the time; nothing further is claimed.
Selected academic activity
- Remote presentation, “Nurturing Chinese Entrepreneurial Spirit in Response to the Belt and Road Initiative,” for Bright Road-Belt and Road: From Strategic Cooperation to Prosperity, held in Almaty, Kazakhstan, 30 November–2 December 2017.
Beyond formal research
Outside formal research, I am interested in a broad range of social phenomena. I observe, write, and think about how behavior, technology, and everyday systems develop, without requiring every observation to fit an academic category. This writing is collected in Notes.