Dongyoon Han

Publications (some preprints are included as well)

• Token Bottleneck: One Token to Remember Dynamics

Taekyung Kim, Dongyoon Han, Byeongho Heo, Jeongeun Park, Sangdoo Yun

Neural Information Processing Systems (NeurIPS), 2025. 

• NegMerge: Consensual Weight Negation for Strong Machine Unlearning
  Hyoseo Kim, Dongyoon Han*, Junsuk Choe* (* corresponding author)
  International Conference on Machine Learning (ICML), 2025.

• Peri-LN: Revisiting Layer Normalization in the Transformer Architecture 

Jeonghoon Kim, Byeongchan Lee, Cheonbok Park, Yeontaek Oh, Beomjun Kim, Taehwan Yoo, Seongjin Shin, Dongyoon Han, Jinwoo Shin, Kang Min Yoo

International Conference on Machine Learning (ICML), 2025.

• Diffusion Meets Few-shot Class Incremental Learning
  Junsu Kim, Yunhoe Ku, Dongyoon Han*, Seungryul Baek* (* corresponding author)
  arXiv:2503.23402, 2025. (under review)

• Masking meets Supervision: A Strong Learning Alliance
  Byeongho Heo, Taekyung Kim, Sangdoo Yun, Dongyoon Han
  Computer Vision and Pattern Recognition (CVPR), 2025.

• Morphing Tokens Draw Strong Masked Image Models

Taekyung Kim*, Byeongho Heo, Dongyoon Han* (* 1st author)
International Conference on Learning Representations (ICLR), 2025.

• DaWin: Training-free Dynamic Weight Interpolation for Robust Adaptation
  Changdae Oh, Sharon Li, Kyungwoo Song*, Sangdoo Yun*, Dongyoon Han* (* corresponding author)
  International Conference on Learning Representations (ICLR), 2025.

• Token-Supervised Value Models for Enhancing Mathematical Problem-Solving Capabilities of Large Language Models

Jung Hyun Lee, June Yong Yang, Byeongho Heo, Dongyoon Han, Kyungsu Kim, Eunho Yang, Kang Min Yoo

International Conference on Learning Representations (ICLR), 2025.

• Self-supervised Visual State Representation Learning for robotics from Dynamic Scenes

Taekyung Kim, Jeongeun Park, Sangdoo Yun, Dongyoon Han, Byeongho Heo

International Conference on Learning Representations (ICLR), 7th Robot Learning Workshop, 2025.

• Why Train Everything? Tint a Single Layer for Multi-task Model Merging
  Aecheon Jung, Seunghwan Lee, Dongyoon Han*, Sungeun Hong* (* corresponding author)
  arXiv:2412.19098, 2025. (under review)

• DNNs May Determine Major Properties of Their Outputs Early, with Timing Possibly Driven by Bias 

Song Park*, Sanghyuk Chun*, Byeongho Heo, Dongyoon Han

arXiv:2502.08167, 2025.  (under review)

• MaskRIS: Semantic Distortion-aware Data Augmentation for Referring Image Segmentation
  Minhyun Lee, Seungho Lee, Song Park, Dongyoon Han, Byeongho Heo, Hyunjung Shim
  arXiv:2411.19067, 2024. (under review,  co-mentored project)

• Token-Supervised Value Models for Enhancing Mathematical Reasoning Capabilities of Large Language Models
  Jung Hyun Lee, June Yong Yang, Byeongho Heo, Dongyoon Han, Kang Min Yoo
  arXiv:2407.12863, 2024.

• Towards Calibrated Robust Fine-Tuning of Vision-Language Models
  Changdae Oh, Hyesu Lim, Mijoo Kim, Dongyoon Han, Sangdoo Yun, Jaegul Choo, Alexander Hauptmann, Zhi-Qi Cheng, Kyungwoo Song
  Neural Information Processing Systems (NeurIPS), 2024. 

• Match Me If You Can: Semi-supervised Semantic Correspondence Learning with Unpaired Images
  Jiwon Kim, Byeongho Heo, Sangdoo Yun, Seungryong Kim, Dongyoon Han* (* corresponding author)
  Asian Conference on Computer Vision (ACCV), 2024. 

• Learning with Unmasked Tokens Drives Stronger Vision Learners
  Taekyung Kim*, Sanghyuk Chun, Byeongho Heo, Dongyoon Han*  (* 1st author)
  European Conference on Computer Vision (ECCV), 2024. 

• Model Stock: All We Need Is Just a Few Fine-Tuned Models
  Dong-Hwan Jang, Sangdoo Yun*, Dongyoon Han* (* corresponding author)
  European Conference on Computer Vision (ECCV), 2024. (oral presentation)

• DenseNets Reloaded: Paradigm Shift Beyond ResNets and ViTs
  Donghyun Kim*, Byeongho Heo, Dongyoon Han* (* 1st author)
  European Conference on Computer Vision (ECCV), 2024.

• Rotary Position Embedding for Vision Transformer
  Byeongho Heo, Song Park, Dongyoon Han, Sangdoo Yun
  European Conference on Computer Vision (ECCV), 2024. 

• HYPE: Hyperbolic Entailment Filtering for Underspecified Images and Texts
  Wonjae Kim, Sanghyuk Chun, Taekyung Kim, Dongyoon Han, Sangdoo Yun
  European Conference on Computer Vision (ECCV), 2024.

• Leveraging Temporal Contextualization for Video Action Recognition
  Minji Kim, Dongyoon Han, Taekyung Kim, Bohyung Han
  European Conference on Computer Vision (ECCV), 2024. (co-mentored project)

• SeiT++: Masked Token Modeling Improves Storage-efficient Training
  Minhyun Lee*, Song Park*, Byeongho Heo, Dongyoon Han, Hyunjung Shim
  European Conference on Computer Vision (ECCV), 2024. (co-mentored project)

• Similarity of Neural Architectures Based on Input Gradient Transferability
  Jaehui Hwang, Dongyoon Han, Byeongho Heo, Song Park, Sanghyuk Chun, Jong-Seok Lee
  European Conference on Computer Vision (ECCV), 2024. (co-mentored project)

• Switching Temporary Teachers for Semi-Supervised Semantic Segmentation
  Jaemin Na, Jung-Woo Ha, Hyung Jin Chang, Dongyoon Han*, Wonjun Hwang* (* corresponding author)
  Neural Information Processing Systems (NeurIPS), 2023.

• Neglected Free Lunch -- Learning Image Classifiers Using Annotation Byproducts
  Dongyoon Han*, Junsuk Choe*, Seonghyeok Chun, John Joon Young Chung, Minsuk Chang, Sangdoo Yun, Jean Y. Song, Seong Joon Oh  (* 1st author)
  International Conference on Computer Vision (ICCV), 2023.

• Gramian Attention Heads are Strong yet Efficient Vision Learners
  Jongbin Ryu*, Dongyoon Han*, Jongwoo Lim (* 1st author)
  International Conference on Computer Vision (ICCV), 2023. 

• Generating Instance-level Prompts for Rehearsal-free Continual Learning
  Dahuin Jung, Dongyoon Han, Jiwhan Bang, Hwanjun Song
  International Conference on Computer Vision (ICCV), 2023. (co-mentored project, oral presentation)

• Scratching Visual Transformer's Back with Uniform Attention
  Nam Hyeon-Woo, Kim Yu-Ji, Byeongho Heo, Dongyoon Han, Seong Joon Oh, Tae-Hyun Oh
  International Conference on Computer Vision (ICCV), 2023. (co-mentored project)

• The Devil is in the Points: Weakly Semi-Supervised Instance Segmentation via Point-Guided Mask Representation
  Beomyoung Kim, Joonhyun Jeong, Dongyoon Han, Sung Ju Hwang
  Conference on Computer Vision and Pattern Recognition (CVPR), 2023.

• TL-ADA: Transferable Loss-based Active Domain Adaptation
  Kyeongtak Han, Youngeun Kim, Dongyoon Han, Sungeun Hong
  Neural Networks, 2023. (co-mentored project)

• Can We Find Strong Lottery Tickets in Generative Models?
  Sangyeop Yeo, Yoojin Jang, Jy-yong Sohn, Dongyoon Han, Jaejun Yoo
  AAAI Conference on Artificial Intelligence (AAAI), 2023. (co-mentored project)

• Spatiotemporal Augmentation on Selective Frequencies for Video Representation Learning
  Jinhyung Kim, Taeoh Kim, Minho Shim, Dongyoon Han, Dongyoon Wee, Junmo Kim
  AAAI Conference on Artificial Intelligence (AAAI), 2023. (co-mentored project)

• Contrastive Vicinal Space for Unsupervised Domain Adaptation
  Jaemin Na, Dongyoon Han, Hyung Jin Chang, and Wonjun Hwang
  European Conference on Computer Vision (ECCV), 2022.

• Donut: Document Understanding Transformer without OCR
  Geewook Kim, Teakgyu Hong, Moonbin Yim, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, and Seunghyun Park
  European Conference on Computer Vision (ECCV), 2022.

• Time Is MattEr: Temporal Self-supervision for Video Transformers
  Sukmin Yun, Jaehyung Kim, Dongyoon Han, Hwanjun Song, Jung-Woo Ha, Jinwoo Shin
  International Conference on Machine Learning (ICML), 2022. (co-mentored project)

• Neural Architecture Search with Loss Flatness-aware Measure
  Joonhyun Jeong, Joonsang Yu, Dongyoon Han, YoungJoon Yoo
  International Conference on Machine Learning (ICML), 2022, Workshop on Dynamic Neural Networks. (co-mentored project)

• Loss-based Sequential Learning for Active Domain Adaptation
  Kyeongtak Han, Youngeun Kim, Dongyoon Han, Sungeun Hong
  arXiv preprint arXiv:2204.11665. (co-mentored project)

• An Extendable, Efficient, and Effective Transformer-based Object Detector
  Hwanjun Song, Deqing Sun, Sanghyuk Chun, Varun Jampani, Dongyoon Han, Byeongho Heo, Wonjae Kim, and Ming-Hsuan Yang
  arXiv preprint arXiv:2204.07962.

• Demystifying the Neural Tangent Kernel from a Practical Perspective: Can it be trusted for Neural Architecture Search without training?
  Jisoo Mok, Byunggook Na, Ji-Hoon Kim*, Dongyoon Han*, Sungroh Yoon* (* corresponding author)
  Conference on Computer Vision and Pattern Recognition (CVPR), 2022. 

• Learning Features with Parameter-Free Layers
  Dongyoon Han, YoungJoon Yoo, Beomyoung Kim, and Byeongho Heo
  International Conference on Learning Representations (ICLR), 2022. (NAVER AI Lab's Best 2022 paper)

• ViDT: An Efficient and Effective Fully Transformer-based Object Detector
  Hwanjun Song, Deqing Sun, Sanghyuk Chun, Varun Jampani, Dongyoon Han, Byeongho Heo, Wonjae Kim, and Ming-Hsuan Yang
  International Conference on Learning Representations (ICLR), 2022.

• Rethinking Spatial Dimensions of Vision Transformers
  Byeongho Heo, Sangdoo Yun, Dongyoon Han, Sanghyuk Chun, Junsuk Choe, and Seong Joon Oh
  International Conference on Computer Vision (ICCV), 2021.

• Detecting and Removing Text in the Wild
  Junho Cho, Sangdoo Yun, Dongyoon Han, Byeongho Heo, and Jin Young Choi
  IEEE Access, vol. 9, pp. 123313-123323, 2021. (co-mentored project)

• Region-based Dropout with Attention Prior for Weakly Supervised Object Localization
  Junsuk Choe, Dongyoon Han, Sangdoo Yun, Jung-Woo Ha, Seong Joon Oh, and Hyunjung Shim
  Pattern Recognition (impact factor: 7.74), 116, 2021.

• Rethinking Channel Dimensions for Efficient Model Design
  Dongyoon Han, Sangdoo Yun, Byeongho Heo, and YoungJoon Yoo
  Conference on Computer Vision and Pattern Recognition (CVPR), 2021.

• Re-labeling ImageNet: From Single to Multi-Labels, from Global to Localized Labels
  Sangdoo Yun, Seong Joon Oh, Byeongho Heo, Dongyoon Han, Junsuk Choe, and Sanghyuk Chun
  Conference on Computer Vision and Pattern Recognition (CVPR), 2021.

• Slowing Down the Weight Norm Increase in Momentum-based Optimizers
  Byeongho Heo*, Sanghyuk Chun*, Seong Joon Oh, Dongyoon Han, Sangdoo Yun, Youngjung Uh, and Jung-Woo Ha
  International Conference on Learning Representations (ICLR), 2021.

• VideoMix: Rethinking Data Augmentation for Video Classification
  Sangdoo Yun, Seong Joon Oh, Byeongho Heo, Dongyoon Han, and Jinhyung Kim
  arXiv preprint arXiv:2012.03457, 2020.

• An Empirical Evaluation on Robustness and Uncertainty of Regularization Methods
  Sanghyuk Chun, Seong Joon Oh, Sangdoo Yun, Dongyoon Han, Junsuk Choe, YoungJoon Yoo
  International Conference on Machine Learning (ICML), 2019, Uncertainty & Robustness in Deep Learning Workshop.

• Unpaired Sketch-to-Line Translation via Synthesis of Sketches
  Gayoung Lee, Dohyun Kim, Youngjoon Yoo, Dongyoon Han, Jung-Woo Ha, Jaehyuk Chang
  SIGGRAPH Asia, 2019. (co-mentored project)

• EXTD: Extremely Tiny Face Detector via Iterative Filter Reuse
  Youngjoon Yoo, Dongyoon Han, and Sangdoo Yun
  arXiv preprint arXiv:1906.0657, 2019.

• CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features
  Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, and Youngjoon Yoo
  International Conference on Computer Vision (ICCV), 2019 (oral presentation)

• What is Wrong with Scene Text Recognition Model Comparisons? Dataset and Model Analysis
  Jeonghun Baek, Geewook Kim, Junyeop Lee, Sungrae Park, Dongyoon Han, Sangdoo Yun, Seong Joon Oh, and Hwalsuk Lee
  International Conference on Computer Vision (ICCV), 2019 (co-mentored project, oral presentation)

• Where to Be Adversarial Perturbations Added? Investigating and Manipulating Pixel Robustness Using Input Gradients
  Jisung Hwang*, Younghoon Kim*, Sanghyuk Chun*, Jaejun Yoo, Ji-Hoon Kim, and Dongyoon Han† († corresponding author)
  International Conference on Learning Representations (ICLR), 2019, Debugging Machine Learning Models Workshop.

• Concentrated-Comprehensive Convolutions for Lightweight Semantic Segmentation
  Hyojin Park, Youngjoon Yoo, Geonseok Seo, Dongyoon Han, Sangdoo Yun, and Nojun Kwak
  arXiv preprint arXiv:1812.04920, 2018. (co-mentored project)

• Character Region Awareness for Text Detection
  Youngmin Baek, Bado Lee, Dongyoon Han, Sangdoo Yun, and Hwalsuk Lee
  IEEE Computer Vision and Pattern Recognition (CVPR), 2019.

• Learning Receptive Field Size by Learning Filter Size
  Yekang Lee, Heechul Jung, Dongyoon Han, Kyungsu Kim, and Junmo Kim
  IEEE Winter Conference on Applications of Computer Vision (WACV), 2019.

• Towards Flatter Loss Surface via Nonmonotonic Learning Rate Scheduling
  Sihyeon Seong, Yekang Lee, Youngwook Kee, Dongyoon Han, and Junmo Kim
  Conference on Uncertainty in Artificial Intelligence (UAI), 2018.

• Unified Simultaneous Clustering and Feature Selection for Unlabeled and Labeled Data
  Dongyoon Han and Junmo Kim
  IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2018 (impact factor: 11.683)

• Deep Pyramidal Residual Networks
  Dongyoon Han*, Jiwhan Kim*, and Junmo Kim (* 1st author)
  IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2017.

• Cost-efficient 3D Face Reconstruction from a Single 2D Image
  Juseung Yun, Jaeyoung Lee, Dongyoon Han, Jeongwoo Ju, and Junmo Kim
  ICACT, 2017. (mentored project)

• 3D Face Recognition via Discriminative Keypoint Selection
  Jiwhan Kim, Dongyoon Han, Wonjun Hwang, and Junmo Kim
  14th International Conference on Ubiquitous Robots and Ambient Intelligence (URAI), 2017.

• Salient Region Detection via High-Dimensional Color Transform and Local Spatial Support
  Jiwhan Kim, Dongyoon Han, Yu-Wing Tai, and Junmo Kim
  IEEE Transactions on Image Processing (TIP), Vol. 25, No. 1, pp. 9-23, 2016 (impact factor: 6.79).

• Facial Age Estimation via Extended Curvature Gabor Filter
  Jiwhan Kim, Dongyoon Han, Sungryull Sohn, and Junmo Kim
  IEEE International Conference on Image Processing (ICIP), 2015.

• Unsupervised Orthogonal Basis Feature Selection
  Dongyoon Han and Junmo Kim
  IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2015.

• Automatic Drawing Simplification via Complex Zernike Moments
  Dongyoon Han, Jung Eun, Pyunghwan Ahn, Jeonghyo Ha, Donghoon Shin, and Junmo Kim
  ITC-CSCC, 2015.

• Salient Region Detection via High Dimensional Color Transform
  Jiwhan Kim, Dongyoon Han, Yu-Wing Tai, and Junmo Kim
  IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2014.

• Efficient and Fast Multi-View Face Detection Based on Feature Transformation
  Dongyoon Han, Jiwhan Kim, Jeongwoo Ju, Injae Lee, Jihun Cha, and Junmo Kim
  ICACT, 2014.

• Texture Classification Based on Discriminative Component Selection of Local Binary Pattern and Variants
  Dongyoon Han and Junmo Kim
  FCV, 2014.

• A Survey of Face Recognition Techniques for Real Media Processing
  Wonjun Hwang, Jiwhan Kim, Dongyoon Han, and Junmo Kim
  Korea Society Broadcast Engineers Magazine, Vol. 19, No. 3, pp. 111-122, Jul. 2014.

Awards

• 4th place at the last ImageNet competition on ILSVRC 2017 object localization task

• Outstanding reviewer at NeurIPS 2018, CVPR 2021

• Outstanding Paper Award at ICACT 2014