Seong Joon Oh

MASEval: Extending Multi-Agent Evaluation from Models to Systems ACL 2026 Privacy Collapse: Benign Fine-Tuning Can Break Contextual Privacy in Language Models ACL 2026 Decoupled Finetuning for Domain Generalizable Semantic Segmentation ICLR 2025 Do Deep Neural Network Solutions Form a Star Domain? ICLR 2025 Does Data Scaling Lead to Visual Compositional Generalization? ICML 2025 Intermediate Layer Classifiers for OOD generalization ICLR 2025 Leaky Thoughts: Large Reasoning Models Are Not Private Thinkers EMNLP 2025 Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models NAACL 2025 Benchmarking Uncertainty Disentanglement: Specialized Uncertainties for Specialized Tasks NeurIPS 2024 TRAP: Targeted Random Adversarial Prompt Honeypot for Black-Box Identification ACL 2024 A Bayesian Approach To Analysing Training Data Attribution In Deep Learning NeurIPS 2023 ID and OOD Performance Are Sometimes Inversely Correlated on Real-world Datasets NeurIPS 2023 Neglected Free Lunch - Learning Image Classifiers Using Annotation Byproducts ICCV 2023 Probabilistic Contrastive Learning Recovers the Correct Aleatoric Uncertainty of Ambiguous Inputs ICML 2023 ProPILE: Probing Privacy Leakage in Large Language Models NeurIPS 2023 Scratching Visual Transformer's Back with Uniform Attention ICCV 2023 URL: A Representation Learning Benchmark for Transferable Uncertainty Estimates NeurIPS 2023 ALP: Data Augmentation Using Lexicalized PCFGs for Few-Shot Text Classification AAAI 2022 Dataset Condensation via Efficient Synthetic-Data Parameterization ICML 2022 ECCV Caption: Correcting False Negatives by Collecting Machine-and-Human-Verified Image-Caption Associations for MS-COCO ECCV 2022 SelecMix: Debiased Learning by Contradicting-pair Sampling NeurIPS 2022 Weakly Supervised Semantic Segmentation Using Out-of-Distribution Data CVPR 2022 Which Shortcut Cues Will DNNs Choose? A Study from the Parameter-Space Perspective ICLR 2022 AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights ICLR 2021 Keep CALM and Improve Visual Feature Attribution ICCV 2021 Neural Hybrid Automata: Learning Dynamics With Multiple Modes and Stochastic Transitions NeurIPS 2021 Probabilistic Embeddings for Cross-Modal Retrieval CVPR 2021 Re-Labeling ImageNet: From Single to Multi-Labels, From Global to Localized Labels CVPR 2021 Rethinking Spatial Dimensions of Vision Transformers ICCV 2021 Evaluating Weakly Supervised Object Localization Methods Right CVPR 2020 Learning De-biased Representations with Biased Representations ICML 2020 Reliable Fidelity and Diversity Metrics for Generative Models ICML 2020 CutMix: Regularization Strategy to Train Strong Classifiers With Localizable Features ICCV 2019 Modeling Uncertainty with Hedged Instance Embeddings ICLR 2019 What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis ICCV 2019 Natural and Effective Obfuscation by Head Inpainting CVPR 2018 Towards Reverse-Engineering Black-Box Neural Networks ICLR 2018 Adversarial Image Perturbation for Privacy Protection -- A Game Theory Perspective ICCV 2017 Exploiting Saliency for Object Segmentation From Image Level Labels CVPR 2017 Generating Descriptions With Grounded and Co-Referenced People CVPR 2017 Person Recognition in Personal Photo Collections ICCV 2015