Profile

News

07. 2023     One paper got accepted to ICCV 2023.
04. 2023     Two papers got accepted to ICML 2023.
02. 2023     Jaehong won the best Ph.D. Dissertation Award from the KAIST College of Engineering!
01. 2023     One paper got accepted to ICLR 2023.
05. 2022     Two papers got accepted to ICML 2022.
01. 2022     Two papers got accepted to ICLR 2022 including one oral presentation (acceptance rate=54/3391=1.6%).


Intro

I'm a Postdoctoral Research Associate in the MURGe-Lab at the University of North Carolina, Chapel Hill, working with Prof. Mohit Bansal. Prior to joining UNC, I received my Ph.D. degree from Korea Advanced Institute of Science and Technology (KAIST) in 2023, advised by Prof. Sung Ju Hwang, and won the best Ph.D. Dissertation Award from the KAIST College of Engineering and School of Computing. During my Ph.D. study, I interned at Microsoft Research, mentored by Dr. Yue Cao, and worked as a short-term visiting student at the Weizmann Institute of Science, hosted by Prof. Yonina Eldar.

My research interest includes following topics:
  • Efficient Deep Learning: Continual Learning, Federated Learning, and Neural Network Compression
  • Multimodal Learning: Egocentric Vision, Video Understanding, Multimodal Deep Generative Models
  • Learning with Real-world Data: Un-(Self-)/Semi-supervised Learning and Input Selective Training

  • My research interest mainly focuses on developing continuously evolving and efficient deep learning alrogithms for deploying sustainable on-device artificial general intelligence systems. In particular, I've been focusing on tackling practical and real-world challenges in application domains, such as online/streaming learning, egocentric videos, and audio-video-text multimodal problems.


    Recent Preprints

    (*: equal contribution)


    concept
    [P3] Analyzing and Mitigating Object Hallucination in Large Vision-Language Models

    Yiyang Zhou*, Chenhang Cui*, Jaehong Yoon, Linjun Zhang, Zhun Deng, Chelsea Finn, Mohit Bansal, Huaxiu Yao

    arXiv:2310.00754, 2023
    Paper

    concept
    [P2] Progressive Fourier Neural Representation for Sequential Video Compilation

    Haeyong Kang, Jaehong Yoon, Dahyun Kim, Sung Ju Hwang, and Chang D. Yoo

    arXiv:2306.11305, 2023
    Paper Code

    concept
    [P1] EVEREST: Efficient Masked Video Autoencoder by Removing Redundant Spatiotemporal Tokens

    Sunil Hwang*, Jaehong Yoon*, Youngwan Lee*, and Sung Ju Hwang

    arXiv:2211.10636, 2022
    Paper Code


    Publications


    concept
    [C13] Text-Conditioned Sampling Framework for Text-to-Image Generation with Masked Generative Models

    Jaewoong Lee*, Sangwon Jang*, Jaehyeong Jo, Jaehong Yoon, Yunji Kim, Jin-Hwa Kim, Jung-Woo Ha, Sung Ju Hwang

    ICCV 2023  
    Project Page Paper

    concept
    [C12] Continual Learners are Incremental Model Generalizers

    Jaehong Yoon, Sung Ju Hwang, and Yue Cao

    ICML 2023  
    Paper

    concept
    [C11] Personalized Subgraph Federated Learning

    Jinheon Baek*, Wonyong Jeong*, Jiongdao Jin, Jaehong Yoon, and Sung Ju Hwang

    ICML 2023  
    Paper Code

    concept
    On-device, Online Continual Learning for the Real World

    Jaehong Yoon

    Ph.D. Thesis   The Best Ph.D. Dissertation Award from the KAIST College of Engineering.
    Paper

    concept
    [C10] On the Soft-Subnetwork for Few-shot Class Incremental Learning

    Haeyong Kang, Jaehong Yoon, Sultan R. H. Madjid, Sung Ju Hwang, and Chang D. Yoo

    ICLR 2023
    Paper Code

    concept
    [W1] BiTAT: Neural Network Binarization with Task-dependent Aggregated Transformation

    Geon Park*, Jaehong Yoon*, Haiyang Zhang, Xing Zhang, Sung Ju Hwang, and Yonina Eldar

    ECCV 2022  Workshop on Computational Aspects of Deep Learning (CADL)
    Paper

    concept
    [C9] Bitwidth Heterogeneous Federated Learning with Progressive Weight Dequantization

    Jaehong Yoon*, Geon Park*, Wonyong Jeong, and Sung Ju Hwang

    ICML 2022
    Paper Code

    concept
    [C8] Forget-free Continual Learning with Winning Subnetworks

    Haeyong Kang*, Rusty J. L. Mina*, Sultan R. H. Madjid, Jaehong Yoon, Mark Hasegawa-Johnson, Sung Ju Hwang, and Chang D. Yoo

    ICML 2022
    Paper Code

    concept
    [C7] Representational Continuity for Unsupervised Continual Learning

    Divyam Madaan, Jaehong Yoon, Yuanchun Li, Yunxin Liu, and Sung Ju Hwang

    ICLR 2022   Oral Presentation (Acceptance Rate=1.6%)
    Paper Code

    concept
    [C6] Online Coreset Selection for Rehearsal-based Continual Learning

    Jaehong Yoon, Divyam Madaan, Eunho Yang, and Sung Ju Hwang

    ICLR 2022
    Paper Code

    concept
    [C5] Federated Continual Learning with Weighted Inter-client Transfer

    Jaehong Yoon*, Wonyong Jeong*, Giwoong Lee, Eunho Yang, and Sung Ju Hwang

    ICML 2020 Workshop on Lifelong Machine Learning
    ICML 2021
    Paper Code

    concept
    [C4] Federated Semi-supervised Learning with Inter-Client Consistency & Disjoint Learning

    Wonyong Jeong, Jaehong Yoon, Eunho Yang, and Sung Ju Hwang

    ICML 2020 Workshop on Federated Learning (Long Presentation) (Best Student Paper Award)
    ICLR 2021
    Paper Code

    concept
    [C3] Scalable and Order-robust Continual Learning with Additive Parameter Decomposition

    Jaehong Yoon, Saehoon Kim, Eunho Yang, and Sung Ju Hwang

    ICLR 2020
    Paper Code

    concept
    [C2] Lifelong Learning with Dynamically Expandable Networks

    Jaehong Yoon, Eunho Yang, Jeongtae Lee, and Sung Ju Hwang

    ICLR 2018
    Paper Code

    concept
    [C1] Combined Group and Exclusive Sparsity for Deep Neural Networks

    Jaehong Yoon and Sung Ju Hwang

    ICML 2017
    Paper Code