Changdae Oh

Changdae Oh

M.S. student at Department of Artificial Intelligence, University of Seoul

About Me

I’m a M.S. student at Machine Learning and Artificial Intelligence (MLAI) lab in University of Seoul, under the supervision of Prof. Jiyoung Jung (adviser) and Prof. Kyungwoo Song (co-adviser). I was fortunate enough to work with Zhi-Qi Cheng, Alexander Hauptmann, and David R. Mortensen during my visiting at Carnegie Mellon University (Aug 23 ~ Feb 24).

For real-world AI applications, I think robustness, reliability, and multimodality are crucial to our intelligence system. To this end, I’ve been working on a robust adaptation of foundation models, multimodal representation learning, and model debiasing & calibration. Topics of interest:

  • Robust and Efficient Adaptation of Foundation Models
  • Debiased Representation Learning
  • Multimodal Learning

News

Apr. 2024. One paper accepted to Expert Systems with Applications
Apr. 2024. I’m starting my internship at NAVER AI Lab
Mar. 2024. I will join UW-Madison CS as a Ph.D. student in 2024f
Feb. 2024. One paper accepted to Pattern Recognition
Oct. 2023. One paper accepted to NeurIPS 2023 Workshop DistShift
Sep. 2023. One paper accepted to NeurIPS 2023
Jun. 2023. One paper accepted to IEEE Access
Mar. 2023. Selected as a Carnegie Mellon University AI Intensive Program Scholarship Recipient (Visiting Scholar, Aug 2023 ~ Feb 2024)
Feb. 2023. One paper accepted to CVPR 2023
Feb. 2023. Outstanding paper award (1st place) at University of Seoul
May. 2022. One paper accepted to KDD 2022

Publication (International)

(conference, journal)

  • Language Model-guided Student Performance Prediction with Multimodal Auxiliary Information
    Changdae Oh, Minhoi Park, Sungjun Lim, Kyungwoo Song
    [paper]
    Expert Systems with Applications 2024

  • Bibimbap: Pre-trained Models Ensemble for Domain Generalization
    Jinho Kang, Taero Kim, Yewon Kim, Changdae Oh, Jiyoung Jung, Rakwoo Chang, Kyungwoo Song
    [paper] [code]
    Pattern Recognition 2024

  • Towards Calibrated Robust Fine-Tuning of Vision-Language Models
    Changdae Oh*, Hyesu Lim*, Mijoo Kim, Jaegul Choo, Alexander Hauptmann, Zhi-Qi Cheng, Kyungwoo Song
    [paper]
    NeurIPS 2023 Workshop, DistShift

  • Geodesic Multi-Modal Mixup for Robust Fine-tuning
    Changdae Oh*, Junhyuk So*, YongTaek Lim, Hoyoon Byun, Minchul Shin, Jong-June Jeon, Kyungwoo Song
    [paper] [code]
    NeurIPS 2023

  • Robust Contrastive Learning with Dynamic Mixed Margin
    [paper] [code]
    Junhyuk So*, YongTaek Lim*, Yewon Kim*, Changdae Oh, Kyungwoo Song
    IEEE Access 2023

  • BlackVIP: Black-Box Visual Prompting for Robust Transfer Learning
    [paper] [code]
    Changdae Oh, Hyeji Hwang, Hee-young Lee, YongTaek Lim, Geunyoung Jung, Jiyoung Jung, Hosik Choi, Kyungwoo Song
    CVPR 2023

  • Learning Fair Representation via Distributional Contrastive Disentanglement
    [paper] [code]
    Changdae Oh, Heeji Won, Junhyuk So, Taero Kim, Yewon Kim, Hosik Choi, Kyungwoo Song
    KDD 2022

Publication (Domestic)

  • Pre-trained Models Ensembling for Domain Generalization in Chemistry Classification
    Jinho Kang, Taero Kim, Yewon Kim, Changdae Oh, Jiyoung Jung, Rakwoo Chang, Kyungwoo Song
    CKAIA 2023

Preprints

  • Mitigating the Linguistic Gap with Phonemic Representations for Robust Multilingual Language Understanding
    [paper]
    Hae Ji Jung, Changdae Oh, Jooeon Kang, Jimin Sohn, Kyungwoo Song, Jinkyu Kim, David R. Mortensen

  • Multimodal Learning for Social Event Analysis
    Changdae Oh, Hoyoon Byun, Minhoi Park, YongTaek Lim, Neil Kim, Kyungwoo Song

  • Perturb-and-Compare Approach for Detecting Out-of-Distribution Samples in Constrained Access Environments
    Hee-young Lee*, Hoyoon Byun*, Changdae Oh, JinYeong Bak, Kyungwoo Song

  • TC-BERT: Large-scale Language Model for Korean Technology Documents
    [paper]
    Hyeji Hwang*, Changdae Oh*, Eunkyeong Lee, Taero Kim, Yewon Kim, Yunjeong Choi, Sungjin Kim, Hosik Choi, Kyungwoo Song

  • Multi-purpose Technology Commercialization Recommender System with Large-scale Korean Language Model
    Hyeji Hwang*, YongTaek Lim*, Changdae Oh*, Seungyeon Kim, Eunkyeong Lee, Yunjeong Choi, Sungjin Kim, Hosik Choi, Kyungwoo Song

  • Graph Perceiver IO: A General Architecture for Graph Structured Data
    [paper]
    Seyun Bae, Hoyoon Byun, Changdae Oh, Yoon-Sik Cho, Kyungwoo Song

(* equal contribution)

Projects

Carnegie Mellon University

University of Seoul

  • Education Contents Relationship Analysis with Multimodal Learning
    • TIPS, Dec. 2022 - Aug. 2023
    • related papers: Language Model-guided Student Performance Prediction with Multimodal Auxiliary Information (preprint)
  • Multimodal Learning for Social Event Analysis
    • HUSTLERS Corp., Oct. 2022 - Dec. 2022
    • related papers: Multimodal Learning for Social Event Analysis (preprint)
  • Multi-purpose Technology Commercialization Documents Recommendation
    • KISTI, Mar. 2022 - Nov. 2022
    • related papers: Multi-purpose Technology Commercialization Recommender System with Large-scale Korean Language Model (preprint)
  • Epidemiological Relevance Evaluation Technology for Vaccination Reactions
  • Keyword Extraction for Technology Commercialization Documents
    • KISTI, June. 2021 - Oct. 2021
    • related papers: TC-BERT: Large-scale Language Model for Korean Technology Documents (preprint)

Awards & Scholarships

  • DEI Scholarship Travel Awards, CVPR, Apr. 2023
  • (Scholarship; USD 41K) AI Intensive Program at Carnegie Mellon University, IITP and Sogang University, Mar. 2023
  • (1st place) Outstanding Paper Award, President’s prize, University of Seoul, Feb. 2023
  • (2nd place) Presentation Award, Workshop on Data-Driven Chemicals Management, University of Seoul, Feb. 2023
  • Student Travel Awards, KDD, Jul. 2022
  • Academic Excellence Scholarship (half-tuition), University of Seoul, Feb. 2021
  • Academic Excellence Scholarship (half-tuition), University of Seoul, Aug. 2020

Academic Services

  • Conference Reviewer
    • 2024: CVPR
  • Conference Volunteer
    • 2022: KDD