I am a fourth-year PhD student in Computer Science at University of Wisconsin-Madison, where I am fortunate to be advised by Frederic Sala. Before that, I was a master’s student at Seoul National University, where I was lucky to learn deep learning, exploratory data analysis, and information theory from Wonjong Rhee. Prior to that, I received B.A in Psychology and B.S. in Computer Science and Engineering from Seoul National University.
My research is focused on foundation models, including large language models and multimodal foundation models. Much of my work aims to efficiently help these models adopt new skills. This involves two prongs: (1) data-centric approaches for obtaining and selecting fine-tuning data, often by using a strategy called weak supervision and (2) efficient adaptation, including training-free approaches like model editing. Specifically, my previous research includes:
Data-centric AI
- Extending weak supervision to any label spaces [ICLR’22]
- Improving fairness of weak supervision [NeurIPS’23]
- Enhancing multimodal data curation via ensemble of curating models [ICCVW’23]
Training-free adaptation
- Robustifying zero-shot models by leveraging insights from language models [NeurIPSW’23, ICLR’24]
- Enhancing data wrangling LLMs with in-context learning [NeurIPSW’23]
- Correcting label distribution mismatch of zero-shot models with optimal transport [Under review]
News
Publications
- OTTER: Improving Zero-Shot Classification via Optimal Transport, Under Review
Changho Shin, Jitian Zhao, Sonia Cromp, Harit Vishwakarma, Frederic Sala
- Zero-Shot Robustification of Zero-Shot Models, ICLR 2024
Also presented in NeurIPS 2023 R0-FoMo Workshop (Best Paper Award Honorable Mention)
Dyah Adila*, Changho Shin*, Linrong Cai, Frederic Sala
- Pool-Search-Demonstrate: Improving Data-wrangling LLMs via better in-context examples, NeurIPS 2023 TRL Workshop
Changho Shin*, Joon Suk Huh*, Elina Choi
- Multimodal Data Curation via Object Detection and Filter Ensembles, ICCV 2023 DataComp Workshop
Changho Shin*, Tzu-heng Huang*, Sui Jiet Tay, Dyah Adila, Frederic Sala
- Mitigating Source Bias for Fairer Weak Supervision, NeurIPS 2023
Changho Shin, Sonia Cromp, Dyah Adila, Frederic Sala
- Can we get smarter than majority vote? Efficient use of individual rater’s labels for content moderation, NeurIPS 2022 ENLSP Workshop
Changho Shin, Alice Schoenauer Sebag
- Universalizing Weak Supervision, ICLR 2022
Changho Shin, Winfred Li, Harit Vishwakarma, Nicholas Roberts, Frederic Sala
- The ENERTALK dataset, 15 Hz electricity consumption data from 22 houses in
Korea, Scientific Data
Changho Shin, Eunjung Lee, Jeongyun Han, Jaeryun Yim, Hyoseop Lee, and Won-
jong Rhee
- Data Requirements for Applying Machine Learning to Energy Disaggregation, Energies
Changho Shin, Seungeun Rho, Hyoseop Lee, and Wonjong Rhee
- Subtask Gated Networks for Non-Intrusive Load Monitoring, AAAI 2019
Changho Shin, Sunghwan Joo, Jaeryun Yim, Hyoseop Lee, Taesup Moon, and Won-
jong Rhee
AWARDS
Work Experience
apartment
Twitter, San Francisco, USA
2022.06 - 2022.08
- Machine Learning Engineer Intern, Health Team
apartment
Encored Technologies, Seoul, Korea
2018.01 - 2020.07
- Data Scientist, Applied Research Team
- Projects: Deep Learning in NILM, Appliance Promotion based on Disaggregated Appliance Usage, Anomaly Detection in Photovoltaic System, Solar Power Generation Prediction
- Researcher, Defense Information Planning Division
- Topics: Informatization Policy, Artificial Intelligence
Teaching Experience
- Teaching assistant for CS 839 (Foundation Models and the Future of Machine Learning), Fall 2023
- Teaching assistant for CS 300 (Programming II), Fall 2022, Spring 2023
- Teaching assistant for CS 760 (Machine Learning), Fall 2021, Spring 2022
- Teaching assistant for CS 320 (Data Programming II), Spring 2021
- Teaching assistant for CS 220 (Data Programming I), Fall 2020