3DB: A Framework for Debugging Computer Vision Models
Guillaume Leclerc, Hadi Salman, Andrew Ilyas, Sai Vemprala, Logan Engstrom, Vibhav Vineet, Kai Yuanqing Xiao, Pengchuan Zhang, Shibani Santurkar, Greg Yang, Ashish Kapoor, Aleksander Madry
Abstract
We introduce 3DB: an extendable, uni ed framework for testing and debugging vision models using photorealistic simulation. We demonstrate, through a wide range of use cases, that 3DB allows users to discover vulnerabilities in computer vision systems and gain insights into how models make decisions. 3DB captures and generalizes many robustness analyses from prior work, and enables one to study their interplay. Finally, we nd that the insights generated by the system transfer to the physical world. We are releasing 3DB as a library 1 alongside a set of example analyses 2 , guides 3 , and documentation 4 . * Work partially completed while at Microsoft Research. † Equal contribution.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Cited by top-tier papers15
- Salient ImageNet: How to discover spurious features in Deep Learning?Sahil Singla, Soheil FeiziICLR 2022 · 144 citations
- 3D Common Corruptions and Data AugmentationOguzhan Fatih Kar, Teresa Yeo, Andrei Atanov, Amir ZamirCVPR 2022 · 80 citations
- Datamodels: Understanding Predictions with Data and Data with PredictionsAndrew Ilyas, Sung Min Park, Logan Engstrom, Guillaume Leclerc et al.ICML 2022 · 66 citations
- Adaptive Testing of Computer Vision ModelsIrena Gao, Gabriel Ilharco, Scott M. Lundberg, Marco Túlio RibeiroICCV 2023 · 49 citations
- Missingness Bias in Model DebuggingSaachi Jain, Hadi Salman, Eric Wong, Pengchuan Zhang et al.ICLR 2022 · 45 citations
Builds on12
- Habitat: A Platform for Embodied AI ResearchManolis Savva, Jitendra Malik, Devi Parikh, Dhruv Batra et al.ICCV 2019 · 1,863 citations
- Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene UnderstandingMike Roberts, Jason Ramapuram, Anurag Ranjan, Atulit Kumar et al.ICCV 2021 · 633 citations
- Noise or Signal: The Role of Image Backgrounds in Object RecognitionKai Yuanqing Xiao, Logan Engstrom, Andrew Ilyas, Aleksander MadryICLR 2021 · 451 citations
- On Completeness-aware Concept-Based Explanations in Deep Neural NetworksChih-Kuan Yeh, Been Kim, Sercan Ömer Arik, Chun-Liang Li et al.NeurIPS 2020 · 390 citations
- Meta-Sim: Learning to Generate Synthetic DatasetsAmlan Kar, Aayush Prakash, Ming-Yu Liu, Eric Cameracci et al.ICCV 2019 · 272 citations
Related papers
- Unadversarial Examples: Designing Objects for Robust VisionHadi Salman, Andrew Ilyas, Logan Engstrom, Sai Vemprala et al.NeurIPS 2021 · 65 citations
- RoboMD: Uncovering Robot Vulnerabilities through Semantic Potential FieldsSom Sagar, Jiafei Duan, Sreevishakh Vasudevan, Yifan Zhou et al.ICLR 2026 · 7 citations
- Universal 3-Dimensional Perturbations for Black-Box Attacks on Video Recognition SystemsShangyu Xie, Han Wang, Yu Kong, Yuan HongS&P 2022 · 32 citations
- LIBERO-Plus: A Progressive Robustness Benchmark for Visual-Language-Action ModelsSenyu Fei, Siyin Wang, Junhao Shi, Zihao Dai et al.CVPR 2026
- Robo3D: Towards Robust and Reliable 3D Perception against CorruptionsLingdong Kong, Youquan Liu, Xin Li, Runnan Chen et al.ICCV 2023 · 151 citations
