Sangwoo Mo
swmo [at] umich [dot] edu
Google Scholar / Github / CV / Twitter
I am a postdoctoral fellow at the University of Michigan, working with Prof. Stella X. Yu. Before joining UMich, I earned my Ph.D. at KAIST under the supervision of Prof. Jinwoo Shin. During my Ph.D., I had the chance to intern or collaborate with various industry and academic labs, including Google AI, Meta AI, NVIDIA AI, Naver AI, and Kakao Brain, where I had the privilege of working with many outstanding mentors. I also received several student fellowships, including the ones from Samsung, Qualcomm, and Korean AI Association.
My research goal is to build the foundation for AI. To this end, I worked on perception (representation learning), creation (generative models), and action (robotics & decision making), within the domains of vision, language, and multimodality. Specifically, my previous research includes:
▪︎ Representation learning
- Vision: NeurIPS'20, NeurIPS'21, CVPRW'22, ICLR'23, ICLR'24
- Language/Multimodal: AAAI'21, NeurIPS'23, CVPR'24
- Analysis (by lottery ticket): ICLR'20, ICLR'21
▪︎ Generative models
- Vision: ICLR'19, NeurIPS'19, CVPRW'20, ICLR'22, TMLR'23
- Language/Multimodal: ICLR'24, ICLR'24, arXiv'24
- Structured: ICMLW'21, NeurIPS'23
▪︎ Robotics & decision making
News
▪︎ 2024/03: I served as a reviewer at ICML, ECCV, TPAMI, and TNNLS.
▪︎ 2024/02: Our paper on language for vision (highlight) is accepted at CVPR'24.
▪︎ 2024/02: I gave a talk at Hanyang University about foundation models for vision, language, and multimodality (slide).
▪︎ 2024/01: Our papers on visual representation (spotlight), LLM long context, and LLM reasoning are accepted at ICLR'24.
▪︎ 2023/11: I served as a reviewer at ICLR, CVPR, and TMLR.
▪︎ 2023/09: Our papers on vision-language model and structured prediction are accepted at NeurIPS'23.
▪︎ 2023/09: I started a postdoctoral fellowship at the University of Michigan.
▪︎ 2023/08: I gave a talk at Korea University about recent trends in visual representation learning (slide).
▪︎ 2023/06: I successfully defended my PhD thesis! 🧑🎓
Publications
C: conference, W: workshop, J: journal, P: preprint / * equal contribution
[P1] GPT Shortcuts: Learning Reusable Functions from User Dialogues
Hyungyu Shin, Yoonjoo Lee, Kwon Ko, Yumin Cho, Jinho Son, Sangwoo Mo, Juho Kim
Under Review
[C16] Discovering and Mitigating Visual Biases through Keyword Explanation
Younghyun Kim*, Sangwoo Mo*, Minkyu Kim, Kyungmin Lee, Jaeho Lee, Jinwoo Shin
CVPR 2024 (highlight) | paper / slide / poster / code
[C15] Learning Hierarchical Image Segmentation For Recognition and By Recognition
Tsung-Wei Ke*, Sangwoo Mo*, Stella X. Yu
ICLR 2024 (spotlight) | paper / slide / poster / code
[C14] SuRe: Summarizing Retrievals using Answer Candidates for Open-domain QA of LLMs
Jaehyung Kim, Jaehyun Nam, Sangwoo Mo, Jongjin Park, Sang-Woo Lee, Minjoon Seo, Jung-Woo Ha, Jinwoo Shin
ICLR 2024 | paper / slide / poster / code
[C13] Hierarchical Context Merging: Better Long Context Understanding for Pre-trained LLMs
Woomin Song*, Seunghyuk Oh*, Sangwoo Mo, Jaehyung Kim, Sukmin Yun, Jung-Woo Ha, Jinwoo Shin
ICLR 2024 | paper / slide / poster / code
[C12] S-CLIP: Semi-supervised Vision-Language Learning using Few Specialist Captions
Sangwoo Mo, Minkyu Kim, Kyungmin Lee, Jinwoo Shin
NeurIPS 2023 | paper / slide / poster / code
[C11] Diffusion Probabilistic Models for Structured Node Classification
Hyosoon Jang, Seonghyun Park, Sangwoo Mo, Sungsoo Ahn
[J2] Breaking the Spurious Causality of Conditional Generation via Fairness Intervention with Corrective Sampling
Junhyun Nam, Sangwoo Mo, Jaeho Lee, Jinwoo Shin
TMLR 2023 | paper / slide / poster / code
[C10] RoPAWS: Robust Semi-supervised Representation Learning from Uncurated Data
Sangwoo Mo, Jong-Chyi Su, Chih-Yao Ma, Mahmoud Assran, Ishan Misra, Licheng Yu, Sean Bell
ICLR 2023 | paper / slide / poster / code
[W3] OAMixer: Object-aware Mixing Layer for Vision Transformers
Hyunwoo Kang*, Sangwoo Mo*, Jinwoo Shin
CVPRW 2022 (Transformers for Vision) | paper / poster / code
▸ Best Paper Award, Korean AI Association Fall Conference, 2022
[C9] Generating Videos with Dynamics-aware Implicit Generative Adversarial Networks
Sihyun Yu*, Jihoon Tack*, Sangwoo Mo*, Hyunsu Kim, Junho Kim, Jung-Woo Ha, Jinwoo Shin
ICLR 2022 | paper / demo / slide / poster / code
[C8] Object-aware Contrastive Learning for Debiased Scene Representation
Sangwoo Mo*, Hyunwoo Kang*, Kihyuk Sohn, Chun-Liang Li, Jinwoo Shin
NeurIPS 2021 | paper / slide / poster / code
▸ Qualcomm Innovation Fellowship Korea, 2021
[W2] Abstract Reasoning via Logic-guided Generation
Sihyun Yu, Sangwoo Mo, Sungsoo Ahn, Jinwoo Shin
ICMLW 2021 (Self-SL for Reasoning and Perception) | paper / slide
[C7] Layer-adaptive Sparsity for the Magnitude-based Pruning
Jaeho Lee, Sejun Park, Sangwoo Mo, Sungsoo Ahn, Jinwoo Shin
ICLR 2021 | paper / slide / code
[C6] MASKER: Masked Keyword Regularization for Reliable Text Classification
Seung-Jun Moon*, Sangwoo Mo*, Kimin Lee, Jaeho Lee, Jinwoo Shin
AAAI 2021 | paper / slide / poster / code
[J1] Deep Neural Network Based Electrical Impedance Tomographic Sensing Methodology for Large-Area Robotic Tactile Sensing
Hyunkyu Park, Kyungseo Park, Sangwoo Mo, Jung Kim
T-RO 2021 | paper
[C5] CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances
Jihoon Tack*, Sangwoo Mo*, Jongheon Jeong, Jinwoo Shin
NeurIPS 2020 | paper / slide / poster / code
▸ Qualcomm Innovation Fellowship Korea, 2020
[W1] Freeze the Discriminator: a Simple Baseline for Fine-Tuning GANs
Sangwoo Mo, Minsu Cho, Jinwoo Shin
CVPRW 2020 (AI for Content Creation) | paper / slide / code
[C4] Lookahead: A Far-sighted Alternative of Magnitude-based Pruning
Sejun Park*, Jaeho Lee*, Sangwoo Mo, Jinwoo Shin
ICLR 2020 | paper / slide / code
[C3] Mining GOLD Samples for Conditional GANs
Sangwoo Mo, Chiheon Kim, Sungwoong Kim, Minsu Cho, Jinwoo Shin
NeurIPS 2019 | paper / slide / poster / code
[C2] Deep Neural Network Approach in Electrical Impedance Tomography-Based Real-Time Soft Tactile Sensor
Hyunkyu Park, Hyosang Lee, Kyungseo Park, Sangwoo Mo, Jung Kim
[C1] InstaGAN: Instance-aware Image-to-Image Translation
Sangwoo Mo, Minsu Cho, Jinwoo Shin
ICLR 2019 | paper / slide / poster / code
▸ Bronze Prize, Samsung Humantech Paper Awards, 2019
Work Experience
Research Experience
▪︎ U Michigan, Research Fellow (host: Prof. Stella X. Yu) / Ann Arbor, MI / Sep 2023 - current
▪︎ NVIDIA AI, Research Intern (host: Dr. Weili Nie) / Santa Clara, CA (remote) / Oct 2022 - Jan 2023
▪︎ Meta AI, Research Intern (host: Dr. Jong-Chyi Su) / Menlo Park, CA / Jun 2022 - Sep 2022
▪︎ Naver AI, Collaborator (host: Dr. Jung-Woo Ha) / Jeongja, South Korea / Jun 2021 - Sep 2021
▪︎ Google AI, Collaborator (host: Dr. Kihyuk Sohn) / Sunnyvale, CA (remote) / Nov 2020 - May 2021
▪︎ Kakao Brain, Research Intern (host: Dr. Sungwoong Kim) / Pangyo, South Korea / Feb 2019 - May 2019
▪︎ POSTECH, Collaborator (host: Prof. Minsu Cho) / Pohang, South Korea (remote) / Jul 2018 - Feb 2020
Business Experience
▪︎ Boston Consulting Group, Research Assistant / Seoul, South Korea / Sep 2014 - Oct 2014
▪︎ Fast Track Asia (venture capital), Business Intern / Seoul, South Korea / Jun 2014 - Aug 2014
Education
▪︎ KAIST, M.S./Ph.D. in Electrical Engineering (advisor: Prof. Jinwoo Shin) / Daejeon, South Korea / Aug 2023
▪︎ POSTECH, B.S. in Mathematics & Industrial Engineering (minor) - summa cum laude / Pohang, South Korea / Feb 2016
▪︎ Hansung Science High School / Seoul, South Korea / Feb 2011
Mentoring
▪︎ Zilin Wang (PhD @ UMich). Co-advised a project, working in progress.
▪︎ Woomin Song (PhD @ KAIST). Co-advised a project, paper accepted at ICLR'24.
▪︎ Younghyun Kim (MS @ KAIST). Co-advised a project, paper accepted at CVPR'24.
▪︎ Hyunwoo Kang (MS @ KAIST). Co-advised projects, papers accepted at NeurIPS'21 and CVPRW'22.
▪︎ Seung-Jun Moon (MS @ KAIST). Co-advised a project, paper accepted at AAAI'21.
Honors & Awards
Research Honors & Awards
▪︎ Bronze Prize, Samsung Humantech Paper Awards, 2019 ($5,000)
▪︎ Qualcomm Innovation Fellowship Korea, 2020 ($4,000)
▪︎ Qualcomm Innovation Fellowship Korea, 2021 ($4,000)
▪︎ Best Paper Award, Korean AI Association Fall Conference, 2022 ($2,000)
▪︎ Reviewer Award: ICML'20, NeurIPS'20, ICLR'21, ICML'21
Business Honors & Awards
▪︎ 1st Prize, Nomura Consulting Case Competition, 2013
▪︎ 1st Prize, Maeil Business Newspaper Case Competition, 2012
▪︎ 1st Prize, GE-McKinsey Leadership Workshop Idea Competition, 2014
▪︎ 23th President of Postech Management Strategy Club (MSSA)
▪︎ Senior member of Young Engineers Honor Society (YEHS)
▪︎ Visiting student at UC Berkeley and Peking university
Selected Talks
▪︎ Hanyang Uni., AI Dept., "Foundation Models for Vision, Language, and Multimodality," Feb 2024
▪︎ U Michigan, CV Group, "Learning Visual Representations from Uncurated Data," Sep 2023
▪︎ Korea Uni., Stats Dept., "Learning Visual Representations from Uncurated Data," Aug 2023
▪︎ KIAS, AI Group, "Robust Semi-supervised Learning from Uncurated Data,'' Mar 2023
▪︎ UCSD, Wang Lab, "Learning Visual Representations from Uncurated Data," Dec 2022
▪︎ Stanford, Wu Lab, "Learning Visual Representations from Uncurated Data," Nov 2022
▪︎ POSTECH, EffL Lab, "Robust Semi-supervised Learning from Uncurated Data," Oct 2022
▪︎ UNIST, AI Dept., "Object-aware Contrastive Learning for Debiased Scene Representation," Nov 2021
Services
▪︎ Conference Reviewer: NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, AAAI
▪︎ Journal Reviewer: TPAMI, TMLR, TNNLS, IJCV, TCSVT, ML
▪︎ Workshop Reviewer: NeurIPS-DGM, CVPR-AICC, ICML-TEACH, NeurIPS-DistShift, CVPR-POETS
Projects
▪︎ Bosch, "Fine-grained Open-world Category Discovery," 2024
▪︎ ADD, "XAI Method for Object Detection in Remote Sensing," 2020-2022
▪︎ IITP, "Continual and Transfer Learning of Generative Models," 2019-2020
▪︎ ETRI, "Anomaly Detection for Time-Series Data from Smart Grids," 2017-2018
▪︎ WiderPlanet, "Online Ads Recommendation Using Deep Contextual Bandits," 2016