Entity Resolution and Deduplication
When do two records describe the same real-world thing — and how do we make that decision auditable at scale?
From fragmented data to meaningful structure
Xinmin Chu
I am a computational science researcher and builder working on entity resolution, deduplication, multiview learning, and evidence-aware knowledge systems. Beyond formal research, I write about institutions, freedom, animals, travel, rhythm, and the details of ordinary life that resist easy classification.
Research, building, and close observation—across data, institutions, and everyday life.
Research
Making merged data answerable for itself
My research asks when two records describe the same real-world thing, how several views of the same data should be combined, and how the resulting decisions — thresholds, blockings, clusterings — can be audited and reproduced rather than taken on faith.
Record Linkage and Blocking
How can candidate pairs be generated so that linkage stays tractable without silently discarding true matches?
Multiview Representation and Fusion
How should multiple views of the same records — lexical, numeric, embedding, tensor — be combined without letting one view dominate or contradict the rest?
Similarity Graphs and Graph Clustering
Once pairwise similarities exist, how should entities be formed — and which clustering assumptions quietly change the answer?
Projects
Research-informed software
Active Research Program ERBOT
An R-based entity-resolution research program spanning multiview representation, scalable blocking, threshold selection, clustering, and evaluation integrity.
Private Mobile Beta KnowledgeTree
A private mobile beta exploring how rough ideas can grow into inspectable, evidence-aware knowledge structures while preserving provenance, uncertainty, and human judgment.
Experimental Prototype DreamReality
An experimental prototype for capturing dream narratives and transforming them into structured storyboards and visual media while preserving the relationship between the original account and generated artifacts.
Beyond the lab
Close observation
Research is one way of paying attention. I also observe animals, institutions, cities, landscapes, and rhythm—subjects that rarely fit into a single discipline but often reveal how systems actually behave.
Writing on these subjects will live in Notes, organized around society and freedom, animals and everyday life, places and distance, rhythm, and poems and fragments.
Elsewhere