2000 character limit reached
Projection Regret: Reducing Background Bias for Novelty Detection via Diffusion Models (2312.02615v1)
Published 5 Dec 2023 in cs.LG and cs.CV
Abstract: Novelty detection is a fundamental task of machine learning which aims to detect abnormal ($\textit{i.e.}$ out-of-distribution (OOD)) samples. Since diffusion models have recently emerged as the de facto standard generative framework with surprising generation results, novelty detection via diffusion models has also gained much attention. Recent methods have mainly utilized the reconstruction property of in-distribution samples. However, they often suffer from detecting OOD samples that share similar background information to the in-distribution data. Based on our observation that diffusion models can \emph{project} any sample to an in-distribution sample with similar background information, we propose \emph{Projection Regret (PR)}, an efficient novelty detection method that mitigates the bias of non-semantic information. To be specific, PR computes the perceptual distance between the test image and its diffusion-based projection to detect abnormality. Since the perceptual distance often fails to capture semantic changes when the background information is dominant, we cancel out the background bias by comparing it against recursive projections. Extensive experiments demonstrate that PR outperforms the prior art of generative-model-based novelty detection methods by a significant margin.
- A survey of outlier detection methodologies. Artificial intelligence review, 22:85–126, 2004.
- Explaining and harnessing adversarial examples. In ICLR, 2015.
- Efficient out-of-distribution detection in digital pathology using multi-head convolutional neural networks. In MIDL, pages 465–478, 2020.
- Out-of-distribution detection for automotive perception. In 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), pages 2938–2943. IEEE, 2021.
- Codit: Conformal out-of-distribution detection in time-series data. arXiv preprint arXiv:2207.11769, 2022.
- C. M. Bishop. Novelty detection and neural network validation. In ICANN, 1994.
- Input complexity and out-of-distribution detection with likelihood-based generative models. In ICLR, 2020.
- Likelihood regret: An out-of-distribution detection score for variational autoencoder. In NeurIPS, 2020.
- Adversarially learned anomaly detection. In ICDM, 2018.
- Backpropagated gradient representations for anomaly detection. In ECCV, 2020.
- High-resolution image synthesis with latent diffusion models. In CVPR, 2022.
- Re-imagen:retrieval-augmented text-to-image generator. In ICLR, 2023.
- Language models are few-shot learners. arXiv preprint arXiv:2005.14165, 2020.
- Adversarial perturbations fool deepfake detectors. In IJCNN, 2020.
- Addressing semantic drift in question generation for semi-supervised question answering. In EMNLP, 2019.
- Image super-resolution via iterative refinement. arXiv preprint arXiv:2104.07636, 2021.
- Multiscale score matching for out-of-distribution detection. In ICLR, 2021.
- Unsupervised out-of-distribution detection with diffusion inpainting. arXiv preprint arXiv:2302.10326, 2023.
- Implicit generation and generalization in energy-based models. In NeurIPS, 2019.
- The unreasonable effectiveness of deep features as a perceptual metric. In CVPR, 2018.
- Consistency models. In ICML, 2023.
- Improved contrastive divergence training of energy-based models. In ICML, 2021.
- Vaebm: A symbiosis between variational autoencoders and energy-based models. In ICLR, 2021.
- Guiding energy-based models via contrastive latent variables. In ICLR, 2023.
- Nvae: A deep hierarchical variational autoencoder. In NeurIPS, 2020.
- Generative modeling by estimating gradients of the data distribution. In NeurIPS, 2019.
- Image quality assessment: From error visibility to structural similarity. IEEE Transactions on Image Processing, 2004.
- Score-based generative modeling through stochastic differential equations. In ICLR, 2021a.
- Elucidating the design space of diffusion-based generative models. In NeurIPS, 2022.
- Learning multiple layers of features from tiny images. Master’s thesis, Department of Computer Science, University of Toronto, 2009.
- Reading digits in natural images with unsupervised feature learning. In NIPS Workshop on Deep Learning and Unsupervised Feature Learning, 2011.
- Unsupervised representation learning by predicting image rotations. In ICLR, 2018.
- Do deep generative models know what they don’t know? In ICLR, 2019.
- Enhancing the reliability of out-of-distribution image detection in neural networks. In ICLR, 2018.
- A simple unified framework for detecting out-of-distribution samples and adversarial attacks. In NeurIPS, 2018.
- Lsun: Construction of a large-scale image dataset using deep learning with humans in the loop. arXiv preprint arXiv:1506.03365, 2015.
- Imagenet: A large-scale hierarchical image database. In CVPR, 2009.
- Describing textures in the wild. In CVPR, 2014.
- Csi: Novelty detection via contrastive learning on distributionally shifted instances. In NeurIPS, 2020.
- Invariant risk minimization. arXiv preprint arXiv:1907.02893, 2019.
- Auto-encoding variational bayes. In NeurIPS, 2014.
- Glow: Generative flow with invertible 1x1 convolutions. In NeurIPS, 2018.
- Memorizing normality to detect anomaly: memory-augmented deep autoencoder for unsupervised anomaly detection. In ICCV, 2019.
- Unsupervised anomaly detection with generative adversarial networks to guide marker discovery. In IPMI, 2017.
- Ocgan: One-class novelty detection using gans with constrained latent representations. In CVPR, 2019.
- Drocc: Deep robust one-class classification. In ICML, 2020.
- Pytorch: An imperative style, high-performance deep learning library. In NeurIPS, 2019.
- Pixelcnn++: a pixelcnn implementation with discretized logistic mixture likelihood and other modifications. In ICLR, 2017.
- P-kdgan: Progressive knowledge distillation with gans for one-class novelty detection. In IJCAI, 2020.
- Using self-supervised learning can improve model robustness and uncertainty. In NeurIPS, 2019.
- U-net: Convolutional networks for biomedical image segmentation. In MICCAI, 2015.
- Label-efficient semantic segmentation with diffusion models. In ICLR, 2022.
- Adaptive prototype learning and allocation for few-shot segmentation. In CVPR, 2021.
- Unsupervised representation for semantic segmentation by implicit cycle-attention contrastive learning. In AAAI, 2022.
- On the impact of spurious correlation for out-of-distribution detection. In AAAI, 2022.
- Imagenet-trained cnns are biased towards texture; increasing shape bias improves accuracy and robustness. In ICLR, 2019.
- Denoising diffusion probabilistic models. In NeurIPS, 2020.
- Pseudo numerical methods for diffusion models on manifolds. In ICLR, 2022.
- Genie: Higher-order denoising diffusion solvers. In NeurIPS, 2022.
- Progressive distillation for fast sampling of diffusion models. In ICLR, 2022.
- Denoising diffusion implicit models. In ICLR, 2021b.
- Novelty detection via blurring. In ICLR, 2020.
- Why normalizing flows fail to detect out-of-distribution data. In NeurIPS, 2020.
- Understanding failures in out-of-distribution detection with deep generative models. In ICML, 2021.
- Learning from failure: training debiased classifier from biased classifier. In NeurIPS, 2020.
- On feature learning in the presence of spurious correlations. In NeurIPS, 2022.
- Discover and cure: concept-aware mitigation of spurious correlations. In ICML, 2023.
- Grad-cam: visual explanations from deep networks via gradient-based localization. In ICCV, 2017.
- Dual diffusion implicit bridges for image-to-image translation. In ICLR, 2023.
- Gans trained by a two time-scale update rule converge to a local nash equilibrium. In NeurIPS, 2017.
Collections
Sign up for free to add this paper to one or more collections.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
