Caleb Belth
Welcome! I'm a PhD candidate at the University of Michigan, where
I work with Danai Koutra.
Some of my interests are data mining, unsupervised learning theory, information theory, and philosophy.
My research focuses on developing graph mining methods that are able to operate on insufficient data by
drawing on ideas from information theory and linguistics, such as entropy and insights from language learning.
Current directions include learning in two-class settings, where only positive examples are available (no negative evidence).
My research has contributed methods for identifying subtle patterns in networks that are too infrequent to be discovered by frequency alone,
and for discovering errors and missing information in incomplete knowledge graphs.
Applications of my work include anomaly detection, suspicious behavior discovery, and city/urban planning, including current projects with the City of Detroit on transportation planning.
I am grateful to have been selected for 2020-2023 NSF GRF and NDSEG fellowships. I will be funded by the NSF GRF starting Fall 2020.
I try my best to be responsible with my research, and continually consider whether my work will be constructive or destructive to society.
Prior to Michigan, I received a B.Sc. in Computer Science from Purdue University, where I was fortunate
to work with
Jennifer Neville,
Dan Goldwasser, and
Daisuke Kihara.
Feel free to contact me at cbelth@umich.edu.
In case you are wondering, the name of this website ("quickshift") is a nod to my life-long love of road trips,
driving, and cars in general.
Publications
-
Mining Persistent Activity in Continually Evolving Networks.
Caleb Belth, Xinyi (Carol) Zheng, Danai Koutra
ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), August 2020 (acceptance rate 17%)
[pdf] [code] -
What is Normal, What is Strange, and What is Missing in a Knowledge Graph: Unified Characterization via Inductive Summarization.
Caleb Belth, Xinyi (Carol) Zheng, Jilles Vreeken, Danai Koutra
ACM The Web Conference (WWW), April 2020 (oral presentation, acceptance rate 19%)
[pdf] [code] [video]
-
Personalized Knowledge Graph Summarization: From the Cloud to Your Pocket
Tara Safavi, Caleb Belth, Lukas Faber, Davide Mottin, Emmanuel Muller, Danai Koutra
IEEE International Conference on Data Mining (ICDM), 2019 (acceptance rate: 9%)
[pdf] -
When to Remember Where You Came from: Node Representation Learning in Higher-order Networks
Caleb Belth, Fahad Kamran, Donna Tjandra, Danai Koutra
IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), 2019 (acceptance rate: 15%)
[pdf] [slides]
Also accepted for presentation at the 15th SIGKDD International Workshop on Mining and Learning with Graphs.
