Seong Joon Oh

I'm a Senior Scientist in Machine Learning at the University of Tübingen currently working on Trustworthy AI. I previously did research at Naver AI Lab and Max Planck Institute for Informatics, working with Mario Fritz and Bernt Schiele. My work focuses on building reliable and responsible AI systems.

My research interests span multiple aspects of Trustworthy AI, including uncertainty quantification, out-of-distribution detection, explainability/interpretability, and privacy/security of machine learning systems. I aim to understand the key limitations of current AI systems and develop systematic solutions to make them more reliable and trustworthy. I've published extensively at top computer vision and machine learning venues like CVPR, ICCV, NeurIPS, and ICLR.

In recent years, I've been particularly interested in uncertainty quantification and probabilistic deep learning. I develop methods to make neural networks aware of their own limitations and communicate them effectively to users. I also work on training data attribution and understanding model behaviors under distribution shifts. Through my research, I hope to contribute to making AI systems that we can confidently deploy in real-world applications.

Publications

CLIP Behaves like a Bag-of-Words Model Cross-modally but not Uni-modally

CLIP Behaves like a Bag-of-Words Model Cross-modally but not Uni-modally

Darina Koishigarina, Arnas Uselis, Seong Joon Oh

Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models

Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models

Haritz Puerto, Martin Gubri, Sangdoo Yun, Seong Joon Oh

arXiv.org 2024

Overcoming Domain Limitations in Open-vocabulary Segmentation

Overcoming Domain Limitations in Open-vocabulary Segmentation

Dongjun Hwang, Seong Joon Oh, Junsuk Choe

arXiv.org 2024

Scalable Ensemble Diversification for OOD Generalization and Detection

Scalable Ensemble Diversification for OOD Generalization and Detection

Alexander Rubinstein, Luca Scimeca, Damien Teney, Seong Joon Oh

arXiv.org 2024

Towards User-Focused Research in Training Data Attribution for Human-Centered Explainable AI

Towards User-Focused Research in Training Data Attribution for Human-Centered Explainable AI

Elisa Nguyen, Johannes Bertram, Evgenii Kortukov, Jean Y. Song, Seong Joon Oh

arXiv.org 2024

Studying Large Language Model Behaviors Under Context-Memory Conflicts With Real Documents

Studying Large Language Model Behaviors Under Context-Memory Conflicts With Real Documents

Evgenii Kortukov, Alexander Rubinstein, Elisa Nguyen, Seong Joon Oh

Do Deep Neural Network Solutions Form a Star Domain?

Do Deep Neural Network Solutions Form a Star Domain?

Ankit Sonthalia, Alexander Rubinstein, Ehsan Abbasnejad, Seong Joon Oh

arXiv.org 2024

Calibrating Large Language Models Using Their Generations Only

Calibrating Large Language Models Using Their Generations Only

Dennis Ulmer, Martin Gubri, Hwaran Lee, Sangdoo Yun, Seong Joon Oh

Annual Meeting of the Association for Computational Linguistics 2024

Benchmarking Uncertainty Disentanglement: Specialized Uncertainties for Specialized Tasks

Benchmarking Uncertainty Disentanglement: Specialized Uncertainties for Specialized Tasks

B'alint Mucs'anyi, Michael Kirchhof, Seong Joon Oh

Neural Information Processing Systems 2024

Pretrained Visual Uncertainties

Pretrained Visual Uncertainties

Michael Kirchhof, Mark Collier, Seong Joon Oh, Enkelejda Kasneci

arXiv.org 2024

TRAP: Targeted Random Adversarial Prompt Honeypot for Black-Box Identification

TRAP: Targeted Random Adversarial Prompt Honeypot for Black-Box Identification

Martin Gubri, Dennis Ulmer, Hwaran Lee, Sangdoo Yun, Seong Joon Oh

Annual Meeting of the Association for Computational Linguistics 2024

Mitigating Shortcut Learning with Diffusion Counterfactuals and Diverse Ensembles

Mitigating Shortcut Learning with Diffusion Counterfactuals and Diverse Ensembles

Luca Scimeca, Alexander Rubinstein, Damien Teney, Seong Joon Oh, A. Nicolicioiu, Y. Bengio

Exploring Practitioner Perspectives On Training Data Attribution Explanations

Exploring Practitioner Perspectives On Training Data Attribution Explanations

Elisa Nguyen, Evgenii Kortukov, Jean Y. Song, Seong Joon Oh

arXiv.org 2023

Trustworthy Machine Learning

B'alint Mucs'anyi, Michael Kirchhof, Elisa Nguyen, Alexander Rubinstein, Seong Joon Oh

arXiv.org 2023

URL: A Representation Learning Benchmark for Transferable Uncertainty Estimates

URL: A Representation Learning Benchmark for Transferable Uncertainty Estimates

M. Kirchhof, B'alint Mucs'anyi, Seong Joon Oh, Enkelejda Kasneci

Neural Information Processing Systems 2023

ProPILE: Probing Privacy Leakage in Large Language Models

ProPILE: Probing Privacy Leakage in Large Language Models

Siwon Kim, Sangdoo Yun, Hwaran Lee, Martin Gubri, Sung-Hoon Yoon, Seong Joon Oh

Neural Information Processing Systems 2023

A Bayesian Approach To Analysing Training Data Attribution In Deep Learning

A Bayesian Approach To Analysing Training Data Attribution In Deep Learning

Elisa Nguyen, Minjoon Seo, Seong Joon Oh

Neural Information Processing Systems 2023

Playing repeated games with Large Language Models

Playing repeated games with Large Language Models

Elif Akata, Lion Schulz, Julian Coda-Forno, Seong Joon Oh, M. Bethge, Eric Schulz

arXiv.org 2023

Neglected Free Lunch – Learning Image Classifiers Using Annotation Byproducts

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

IEEE International Conference on Computer Vision 2023

Probabilistic Contrastive Learning Recovers the Correct Aleatoric Uncertainty of Ambiguous Inputs

Probabilistic Contrastive Learning Recovers the Correct Aleatoric Uncertainty of Ambiguous Inputs

M. Kirchhof, Enkelejda Kasneci, Seong Joon Oh

International Conference on Machine Learning 2023

SelecMix: Debiased Learning by Contradicting-pair Sampling

SelecMix: Debiased Learning by Contradicting-pair Sampling

Inwoo Hwang, Sangjun Lee, Yunhyeok Kwak, Seong Joon Oh, Damien Teney, Jin-Hwa Kim, Byoung-Tak Zhang

Neural Information Processing Systems 2022

Scratching Visual Transformer's Back with Uniform Attention

Scratching Visual Transformer's Back with Uniform Attention

Nam Hyeon-Woo, Kim Yu-Ji, Byeongho Heo, Doonyoon Han, Seong Joon Oh, Tae-Hyun Oh

IEEE International Conference on Computer Vision 2022

ID and OOD Performance Are Sometimes Inversely Correlated on Real-world Datasets

Damien Teney, Seong Joon Oh, Ehsan Abbasnejad

Neural Information Processing Systems 2022

Dataset Condensation via Efficient Synthetic-Data Parameterization

Dataset Condensation via Efficient Synthetic-Data Parameterization

Jang-Hyun Kim, Jinuk Kim, Seong Joon Oh, Sangdoo Yun, Hwanjun Song, Joonhyun Jeong, Jung-Woo Ha, Hyun Oh Song

International Conference on Machine Learning 2022

ECCV Caption: Correcting False Negatives by Collecting Machine-and-Human-verified Image-Caption Associations for MS-COCO

ECCV Caption: Correcting False Negatives by Collecting Machine-and-Human-verified Image-Caption Associations for MS-COCO

Sanghyuk Chun, Wonjae Kim, Song Park, Minsuk Chang, Seong Joon Oh

European Conference on Computer Vision 2022

Weakly Supervised Semantic Segmentation using Out-of-Distribution Data

Weakly Supervised Semantic Segmentation using Out-of-Distribution Data

Jungbeom Lee, Seong Joon Oh, Sangdoo Yun, Junsuk Choe, Eunji Kim, Sung-Hoon Yoon

Computer Vision and Pattern Recognition 2022

ALP: Data Augmentation using Lexicalized PCFGs for Few-Shot Text Classification

ALP: Data Augmentation using Lexicalized PCFGs for Few-Shot Text Classification

Hazel Kim, Daecheol Woo, Seong Joon Oh, Jeong-Won Cha, Yo-Sub Han

AAAI Conference on Artificial Intelligence 2021

Which Shortcut Cues Will DNNs Choose? A Study from the Parameter-Space Perspective

Which Shortcut Cues Will DNNs Choose? A Study from the Parameter-Space Perspective

Luca Scimeca, Seong Joon Oh, Sanghyuk Chun, Michael Poli, Sangdoo Yun

International Conference on Learning Representations 2021

Keep CALM and Improve Visual Feature Attribution

Keep CALM and Improve Visual Feature Attribution

Jae Myung Kim, Junsuk Choe, Zeynep Akata, Seong Joon Oh

IEEE International Conference on Computer Vision 2021

Neural Hybrid Automata: Learning Dynamics with Multiple Modes and Stochastic Transitions

Neural Hybrid Automata: Learning Dynamics with Multiple Modes and Stochastic Transitions

Michael Poli, Stefano Massaroli, Luca Scimeca, Seong Joon Oh, Sanghyuk Chun, A. Yamashita, H. Asama, Jinkyoo Park, Animesh Garg

Neural Information Processing Systems 2021

Rethinking Spatial Dimensions of Vision Transformers

Rethinking Spatial Dimensions of Vision Transformers

Byeongho Heo, Sangdoo Yun, Dongyoon Han, Sanghyuk Chun, Junsuk Choe, Seong Joon Oh

IEEE International Conference on Computer Vision 2021

Region-based dropout with attention prior for weakly supervised object localization

Junsuk Choe, Dongyoon Han, Sangdoo Yun, Jung-Woo Ha, Seong Joon Oh, Hyunjung Shim

Pattern Recognition 2021

Re-labeling ImageNet: from Single to Multi-Labels, from Global to Localized Labels

Re-labeling ImageNet: from Single to Multi-Labels, from Global to Localized Labels

Sangdoo Yun, Seong Joon Oh, Byeongho Heo, Dongyoon Han, Junsuk Choe, Sanghyuk Chun

Computer Vision and Pattern Recognition 2021

Probabilistic Embeddings for Cross-Modal Retrieval

Sanghyuk Chun, Seong Joon Oh, Rafael Sampaio de Rezende, Yannis Kalantidis, Diane Larlus

Computer Vision and Pattern Recognition 2021

VideoMix: Rethinking Data Augmentation for Video Classification

VideoMix: Rethinking Data Augmentation for Video Classification

Sangdoo Yun, Seong Joon Oh, Byeongho Heo, Dongyoon Han, Jinhyung Kim

arXiv.org 2020

Evaluation for Weakly Supervised Object Localization: Protocol, Metrics, and Datasets

Evaluation for Weakly Supervised Object Localization: Protocol, Metrics, and Datasets

Junsuk Choe, Seong Joon Oh, Sanghyuk Chun, Zeynep Akata, Hyunjung Shim

IEEE Transactions on Pattern Analysis and Machine Intelligence 2020

AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights

AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights

Byeongho Heo, Sanghyuk Chun, Seong Joon Oh, Dongyoon Han, Sangdoo Yun, Gyuwan Kim, Youngjung Uh, Jung-Woo Ha

International Conference on Learning Representations 2020

Slowing Down the Weight Norm Increase in Momentum-based Optimizers

Slowing Down the Weight Norm Increase in Momentum-based Optimizers

Byeongho Heo, Sanghyuk Chun, Seong Joon Oh, Dongyoon Han, Sangdoo Yun, Youngjung Uh, Jung-Woo Ha

arXiv.org 2020

An Empirical Evaluation on Robustness and Uncertainty of Regularization Methods

An Empirical Evaluation on Robustness and Uncertainty of Regularization Methods

Sanghyuk Chun, Seong Joon Oh, Sangdoo Yun, Dongyoon Han, Junsuk Choe, Y. Yoo

arXiv.org 2020

Reliable Fidelity and Diversity Metrics for Generative Models

Reliable Fidelity and Diversity Metrics for Generative Models

Muhammad Ferjad Naeem, Seong Joon Oh, Youngjung Uh, Yunjey Choi, Jaejun Yoo

International Conference on Machine Learning 2020

Evaluating Weakly Supervised Object Localization Methods Right

Evaluating Weakly Supervised Object Localization Methods Right

Junsuk Choe, Seong Joon Oh, Seungho Lee, Sanghyuk Chun, Zeynep Akata, Hyunjung Shim

Computer Vision and Pattern Recognition 2020

On Recognizing Texts of Arbitrary Shapes with 2D Self-Attention

On Recognizing Texts of Arbitrary Shapes with 2D Self-Attention

Junyeop Lee, Sungrae Park, Jeonghun Baek, Seong Joon Oh, Seonghyeon Kim, Hwalsuk Lee

2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) 2019

Learning De-biased Representations with Biased Representations

Learning De-biased Representations with Biased Representations

Hyojin Bahng, Sanghyuk Chun, Sangdoo Yun, J. Choo, Seong Joon Oh

International Conference on Machine Learning 2019

CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features

CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features

Sangdoo Yun, Dongyoon Han, Seong Joon Oh, Sanghyuk Chun, Junsuk Choe, Y. Yoo

IEEE International Conference on Computer Vision 2019

What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis

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, Hwalsuk Lee

IEEE International Conference on Computer Vision 2019

Modeling Uncertainty with Hedged Instance Embedding

Modeling Uncertainty with Hedged Instance Embedding

Seong Joon Oh, K. Murphy, Jiyan Pan, Joseph Roth, Florian Schroff, Andrew C. Gallagher

International Conference on Learning Representations 2018

Understanding and Controlling User Linkability in Decentralized Learning

Understanding and Controlling User Linkability in Decentralized Learning

Tribhuvanesh Orekondy, Seong Joon Oh, B. Schiele, Mario Fritz

arXiv.org 2018

Gradient-Leaks: Understanding and Controlling Deanonymization in Federated Learning

Gradient-Leaks: Understanding and Controlling Deanonymization in Federated Learning

Tribhuvanesh Orekondy, Seong Joon Oh, Yang Zhang, B. Schiele, Mario Fritz

Sequential Attacks on Agents for Long-Term Adversarial Goals

Sequential Attacks on Agents for Long-Term Adversarial Goals

E. Tretschk, Seong Joon Oh, Mario Fritz

arXiv.org 2018

Natural and Effective Obfuscation by Head Inpainting

Natural and Effective Obfuscation by Head Inpainting

Qianru Sun, Liqian Ma, Seong Joon Oh, L. Gool, B. Schiele, Mario Fritz

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition 2017

Whitening Black-Box Neural Networks

Whitening Black-Box Neural Networks

Seong Joon Oh, Maximilian Augustin, B. Schiele, Mario Fritz

International Conference on Learning Representations 2017

Towards Reverse-Engineering Black-Box Neural Networks

Seong Joon Oh, Maximilian Augustin, Mario Fritz, B. Schiele

International Conference on Learning Representations 2017

Person Recognition in Personal Photo Collections

Person Recognition in Personal Photo Collections

Seong Joon Oh, Rodrigo Benenson, Mario Fritz, B. Schiele

IEEE Transactions on Pattern Analysis and Machine Intelligence 2017

Generating Descriptions with Grounded and Co-referenced People

Generating Descriptions with Grounded and Co-referenced People

Anna Rohrbach, Marcus Rohrbach, Siyu Tang, Seong Joon Oh, B. Schiele

Computer Vision and Pattern Recognition 2017

Adversarial Image Perturbation for Privacy Protection A Game Theory Perspective

Adversarial Image Perturbation for Privacy Protection A Game Theory Perspective

Seong Joon Oh, Mario Fritz, B. Schiele

IEEE International Conference on Computer Vision 2017

Exploiting Saliency for Object Segmentation from Image Level Labels

Exploiting Saliency for Object Segmentation from Image Level Labels

Seong Joon Oh, Rodrigo Benenson, A. Khoreva, Zeynep Akata, Mario Fritz, B. Schiele

Computer Vision and Pattern Recognition 2017

Faceless Person Recognition: Privacy Implications in Social Media

Seong Joon Oh, Rodrigo Benenson, Mario Fritz, B. Schiele

European Conference on Computer Vision 2016

I-Pic: A Platform for Privacy-Compliant Image Capture

I-Pic: A Platform for Privacy-Compliant Image Capture

Paarijaat Aditya, Rijurekha Sen, P. Druschel, Seong Joon Oh, Rodrigo Benenson, Mario Fritz, B. Schiele, Bobby Bhattacharjee, Tongyu Wu

ACM SIGMOBILE International Conference on Mobile Systems, Applications, and Services 2016

Person Recognition in Personal Photo Collections

Person Recognition in Personal Photo Collections

Seong Joon Oh, Rodrigo Benenson, Mario Fritz, B. Schiele

IEEE International Conference on Computer Vision 2015

CYP1A2 activity as a risk factor for bladder cancer.

Seong Won Lee, I. Jang, Sang‐Goo Shin, Kyeong Hoon Lee, Dong-Seok Yim, Si Whang Kim, Seong Joon Oh, Sun-Hee Lee

Journal of Korean medical science 1994

Studying Large Language Model Behaviors Under Realistic Knowledge Conflicts

Studying Large Language Model Behaviors Under Realistic Knowledge Conflicts

Evgenii Kortukov, Alexander Rubinstein, Elisa Nguyen, Seong Joon Oh

arXiv.org 2024

SelecMix: Debiased Learning by Mixing up Contradicting Pairs

SelecMix: Debiased Learning by Mixing up Contradicting Pairs

Inwoo Hwang, Sangjun Lee, Yunhyeok Kwak, Seong Joon Oh, Damien Teney, Jin-Hwa Kim, Byoung-Tak Zhang

Gradient-Leaks: Understanding Deanonymization in Federated Learning

Tribhuvanesh Orekondy, Seong Joon Oh, Yang Zhang, B. Schiele, Mario Fritz

Modeling Uncertainty with Hedged Instance Embeddings

Seong Joon Oh, K. Murphy, Jiyan Pan, Joseph Roth, Florian Schroff, Andrew C. Gallagher

International Conference on Learning Representations 2019

Image manipulation against learned models: privacy and security implications

Image manipulation against learned models: privacy and security implications

Seong Joon Oh

From Understanding to Controlling Privacy against Automatic Person Recognition in Social Media

From Understanding to Controlling Privacy against Automatic Person Recognition in Social Media

Seong Joon Oh, Mario Fritz, B. Schiele, Max

Exercise 3 : Deep Neural Networks and Backpropagation

Exercise 3 : Deep Neural Networks and Backpropagation

Dr. Mario Fritz, Seong Joon Oh, A. Dima

Demo: I-Pic: A Platform for Privacy-Compliant Image Capture

Demo: I-Pic: A Platform for Privacy-Compliant Image Capture

Paarijaat Aditya, Rijurekha Sen, P. Druschel, Seong Joon Oh, Rodrigo Benenson, Mario Fritz, B. Schiele, Bobby Bhattacharjee, Tongyu Wu

MobiSys '16 Companion 2016