Chi-Heng Lin

ECE Ph.D. student

Georgia Institute of Technology

Biography

Chi-Heng is an ECE PhD student in the NerDS Lab at GeorgiaTech, led by Dr. Eva L Dyer. His research experience includes self-supervised learning, Gaussian processes, and optimal transport

Interests

Data augmentation
Domain adaptation
Optimization

Education

PhD in Electrical and Computer Engineering
Georgia Institute of Technology
M.A. in Statistics
Columbia University
B.S. & M.S. in Electrical Engineering
National Taiwan University

Recent Publications

Chi-Heng Lin, Chiraag Kaushik, Eva L. Dyer, Vidya Muthukumar (2022). [Preprint] The good, the bad and the ugly sides of data augmentation: An implicit spectral regularization perspective.

PDF

Ran Liu, Mehdi Azabou, Max Dabagia, Chi-Heng Lin, Mohammad Gheshlaghi Azar, Keith B. Hengen, Michal Valko, Eva L. Dyer (2021). Drop, Swap, and Generate: A Self-Supervised Approach for Generating Neural Activity. NeurIPS 2021 (oral presentation).

PDF Cite

Chi-Heng Lin, Joseph D Miano, Eva L Dyer (2021). Bayesian optimization for modular black-box systems with switching costs. UAI 2021.

PDF Cite Project

Chi-Heng Lin, Mehdi Azabou, Eva L Dyer (2021). Making transport more robust and interpretable by moving data through a small number of anchor points. ICML 2021.

PDF Cite Project

Jun-Kun Wang, Chi-Heng Lin, Jacob Abernethy (2021). A Modular Analysis of Provable Acceleration via Polyak's Momentum: Training a Wide ReLU Network and a Deep Linear Network. ICML 2021.

PDF Cite

See all publications

Projects

Low rank optimal transport

Understand Data Augmentation through Linear Models

Featured Publications

Ran Liu, Mehdi Azabou, Max Dabagia, Chi-Heng Lin, Mohammad Gheshlaghi Azar, Keith B. Hengen, Michal Valko, Eva L. Dyer

June 2021 NeurIPS 2021 (oral presentation)

Drop, Swap, and Generate: A Self-Supervised Approach for Generating Neural Activity

Chi-Heng Lin, Joseph D Miano, Eva L Dyer

May 2021 UAI 2021

Bayesian optimization for modular black-box systems with switching costs

Click the Cite button above to demo the feature to enable visitors to import publication metadata into their reference management software. Create your slides in Markdown - click the Slides button to check out the example.

Chi-Heng Lin, Mehdi Azabou, Eva L Dyer

May 2021 ICML 2021