My research areas are data mining and machine learning. My recent work centers
around acquiring knowledge from unstructured text data, with emphasis on
improving label efficiency and robustness of learning algorithms.
Methodologically, I am particularly interested in weakly-supervised learning
(CIKM’18, AAAI’19), unsupervised
probabilistic models (KDD’19), and uncertainty estimation
Prior to joining Georgia Tech, I obtained my Ph.D. in Computer
Science from UIUC, where I
worked with Jiawei Han on text mining
ICDM’18) and spatiotemporal data mining
For prospective students: We have openings for strong and motivated
students. If you are interested in joining my group, please send me an email
along with your CV and transcript. I may not be able to respond to all the
inquiries due to the large volume of them, but I will contact you if there is a
good match between us.
- 2019.06: “TopicMine: User-Guided Topic Mining by Category-Oriented Embedding” is accepted by KDD 2019 Projects Showcase for oral presentation!
- 2019.04: Our paper “State-sharing Sparse Hidden Markov Models” for modeling sparse sequential data is accepted by KDD 2019 as an oral paper in the research track!
- 2019.04: Our paper “A gradual, semi-discrete approach to generative network training via explicit Wasserstein minimization” is accepted by ICML 2019!
- 2019.02: Our book “Multidimensional Mining of Massive Text Data” has been published by Morgan & Claypool! Check it out here.
- 2018.12: One paper on integrating heterogeneous mobility data is accepted by WWW 2019.
- 2018.12: Two papers are accepted by SDM 2019. One is about using hidden Markov model for mobility modeling; another about cross-modal linking between text and location.
- 2018.11: The code for our CIKM paper on weakly-supervised text classification is released. If you want to use deep neural nets to do text classification but don’t have much labeled data, try it out!
- 2018.11: Our paper on weakly-supervised hierarchical text classification is accepted by AAAI 2019.
- 2018.10: Our paper on linear-time trajectory similarity computation is accepted by ICDE 2019. It’s up to 100x faster than state-of-the-art trajectory similarity computation techniques and supports all major measures! Code will be out soon.
- 2018.08: One paper on weakly-supervised text classification is accepted by CIKM 2018. No need for excessive training data but just a few seeds. Congrats Yu!
- 2018.08: One paper on open-domain event extraction is accepted by CIKM 2018.
- 2018.08: One paper on text cube construction without labeled data is accepted by ICDM 2018.
- 2018.08: Delivered a tutorial and presented three papers in KDD 2018@London.
- 2018.07: Our paper on using GAN for semi-supervised learning is accepted by IMWUT 2018.
- 2018.07: Serving on the Program Committee of AAAI 2019.
- 2018.07: We will give a tutorial on multidimensional analysis of text data (website) in KDD 2018.
- 2018.06: One paper on event detection is accepted by ASONAM 2018.
- 2018.05: Three papers are accepted by KDD 2018 (Research Track).
- 2018.03: Serving on the PC of the GIScience 2018 Workshop “Spatial Big Data and Machine Learning. CFP.
- 2018.02: Our paper “Deep Learning for the Internet of Things” is accepted by IEEE Computer.
- 2018.01: Serving on the Program Committee of IJCAI 2018 and KDD 2018 (Research Track).
- 2017.12: One paper is accepted by WWW 2018.
- 2017.11: Invited to talk about event detection on Illinois Innovator Podcast, available on SoundCloud and iTunes.
- 2017.11: Serving on the Program Committee (Machine Learning Track) of ACL 2018.
- 2017.11: Two papers are accepted by AAAI 2018.
- 2017.10: Check out our Urbanity system (paper) for predicting spatiotemporal activities with social media.
- 2017.10: We give a tutorial on “Space and Time Coupled Social Media Analysis” (slides) in CIKM’17.
- 2017.09: Check out the report from IT Business about our work on event detection from social media.
- 2017.04: We give a tutorial on “Bringing Semantics to Spatiotemporal Data Mining” (slides) in ICDE’17.