Span2_thumb_polo

Duen Horng (Polo) Chau

Alumni

http://www.cs.cmu.edu/~dchau/

Duen Horng "Polo" Chau is a final year Ph.D. student in the Machine Learning Department at Carnegie Mellon University.

Polo's research bridges Data Mining and Human-Computer Interaction to synthesize systems and tools that help people understand and interact with Big Data. His thesis focuses on massive networks with billions of nodes and edges.

He blends techniques from machine learning (Belief Propagation), data mining (anomaly detection), visualization and user interaction. Notable projects:

  • Apolo: mixed-initiative system (ML+vis) for making sense of large network data; Apolo helps users incrementally find relevant subgraphs to explore, avoids overwhelming the user
  • Polonium: patent-pending malware detection technology (with Symantec); infers file reputation over 37 billion machine-file relationships
  • NetProbe: auction fraud detection (eBay) that fingers bad guys by identifying their suspicous transactions; appeared in Wall Street Journal, USA Today, and more
  • PEGASUS: award-winning project that creates graph algorithms for massive graph data

Polo is also an award-winning designer. His design of the Carnegie Mellon ID card has been in use since 2006.

Publications

Data Mining Meets HCI: Making Sense of Large Graphs Duen Horng (Polo) Chau PhD Thesis Ph.D. Thesis
Making Sense of Large Network Data: Combining Rich User Interaction and Machine Learning Duen Horng (Polo) Chau, Aniket Kittur, Jason Hong, and Christos Faloutsos ACM Conference on Human Factors in Computing Systems (CHI) Full Paper
Apolo: Interactive Large Graph Sense making by Combining Machine Learning and Visualization Duen Horng (Polo) Chau, Aniket Kittur, Jason Hong, and Christos Faloutsos ACM SIGKDD Conference on Knowledge Discovery and Data Mining Demo
Supporting Sensemaking Through Large Scale Graph Mining Duen Horng (Polo) Chau, Aniket Kittur, Christos Faloutsos, and Jason Hong Sensemaking Workshop at ACM Conference on Human Factors in Computing Systems (CHI) Workshop Paper