Scholar
Changho Shin
Google Scholar ID: mF95I24AAAAJ
University of Wisconsin-Madison
Citations & Impact
Citations
450
H-index
7
i10-index
6
Publications
16
Co-authors
21
list available
Publications
Resume (English only)
Academic Achievements
- 1. Paper 'Weak-to-Strong Generalization Through the Data-Centric Lens' accepted at ICLR 2025
- 2. Paper 'Personalize Your LLM: Fake it then Align it' accepted at NAACL 2025 Findings
- 3. Paper 'OTTER: Improving Zero-Shot Classification via Optimal Transport' accepted at NeurIPS 2024
- 4. Paper 'Zero-Shot Robustification of Zero-Shot Models' accepted at ICLR 2024 and presented in NeurIPS 2023 R0-FoMo Workshop (Best Paper Award Honorable Mention)
- 5. Paper 'Mitigating Source Bias for Fairer Weak Supervision' accepted at NeurIPS 2023
- 6. Paper 'Universalizing Weak Supervision' accepted at ICLR 2022
- 7. Paper 'Subtask Gated Networks for Non-Intrusive Load Monitoring' accepted at AAAI 2019
- 8. Journal paper 'The ENERTALK dataset, 15 Hz electricity consumption data from 22 houses in Korea' published in Scientific Data
- 9. Journal paper 'Data Requirements for Applying Machine Learning to Energy Disaggregation' published in Energies
- 10. Workshop paper 'Curriculum Learning as Transport: Training Along Wasserstein Geodesics' accepted at NeurIPS 2025 CCFM Workshop
Research Experience
- 1. Summer internship at MSR New England in 2025
- 2. Summer internship at Snorkel AI in 2024
- 3. Team with Dyah Adila shortlisted as finalists in Qualcomm Innovation Fellowship 2024
Education
- 1. Ph.D. in Computer Science at University of Wisconsin-Madison, Advisor: Frederic Sala
- 2. Master's student at Seoul National University, Advisor: Wonjong Rhee
- 3. B.A. in Psychology and B.S. in Computer Science and Engineering from Seoul National University
Background
- Ph.D. student in Computer Science focusing on data-centric AI, programmatic weak supervision, and weak-to-strong generalization in foundation models. Research also includes inference-time intervention to improve robustness, alignment, and personalization of foundation models.
Miscellany
- Email: cshin23@wisc.edu
- Twitter and Github accounts available
