mlsanityData to 14 Oct 2026

Backpropagated Gradient Representations for Anomaly Detection

arXiv 2007.09507 ↗First posted 18 Jul 2020cs.CVcited by 54

Not on the map. Why?

Abstract (arXiv)

Learning representations that clearly distinguish between normal and abnormal data is key to the success of anomaly detection. Most of existing anomaly detection algorithms use activation representations from forward propagation while not exploiting gradients from backpropagation to characterize data. Gradients capture model updates required to represent data. Anomalies require more drastic model updates to fully represent them compared to normal data. Hence, we propose the utilization of backpropagated gradients as representations to characterize model behavior on anomalies and, consequently, detect such anomalies. We show that the proposed method using gradient-based representations achieves state-of-the-art anomaly detection performance in benchmark image recognition datasets. Also, we highlight the computational efficiency and the simplicity of the proposed method in comparison with other state-of-the-art methods relying on adversarial networks or autoregressive models, which require at least 27 times more model parameters than the proposed method.

Key facts. Cited by 54 papers in the map (data through 14 Oct 2026). First posted on arXiv 18 Jul 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 10; 2022: 13; 2023: 16; 2024: 11; 2025: 3; 2026: 1.

54 papers.

Sort:

By area:

  1. 2021Generalized Out-of-Distribution Detection: A Surveycited by 802
  2. 2021Open-Set Recognition: a Good Closed-Set Classifier is All You Need?cited by 353
  3. 2022A Unified Model for Multi-class Anomaly Detectioncited by 268
  4. 2022Omni-frequency Channel-selection Representations for Unsupervised Anomaly Detectioncited by 80
  5. 2021Deep One-Class Classification via Interpolated Gaussian Descriptorcited by 67
  6. 2023Exploring Plain ViT Reconstruction for Multi-class Unsupervised Anomaly Detectioncited by 47
  7. 2023Rethinking Out-of-distribution (OOD) Detection: Masked Image Modeling is All You Needcited by 46
  8. 2022ADTR: Anomaly Detection Transformer with Feature Reconstructioncited by 31
  9. 2022Introspective Learning : A Two-Stage Approach for Inference in Neural Networkscited by 20
  10. 2022Gradient-Based Adversarial and Out-of-Distribution Detectioncited by 14
  11. 2024CLIPScope: Enhancing Zero-Shot OOD Detection with Bayesian Scoringcited by 13
  12. 2021Extracting Causal Visual Features for Limited label Classificationcited by 12
38 more
  1. 2021Contrastive Reasoning in Neural Networkscited by 11
  2. 2022Explanatory Paradigms in Neural Networkscited by 11
  3. 2023Gaussian Switch Sampling: A Second Order Approach to Active Learningcited by 10
  4. 2023Probing the Purview of Neural Networks via Gradient Analysiscited by 9
  5. 2023Projection Regret: Reducing Background Bias for Novelty Detection via Diffusion Modelscited by 9
  6. 2023Forgetful Active Learning with Switch Events: Efficient Sampling for Out-of-Distribution Datacited by 8
  7. 2022Open-Set Recognition with Gradient-Based Representationscited by 7
  8. 2022Man-recon: manifold learning for reconstruction with deep autoencoder for smart seismic interpretationcited by 7
  9. 2023Unveiling the Unseen: A Comprehensive Survey on Explainable Anomaly Detection in Images and Videoscited by 7
  10. 2021Elsa: Energy-based learning for semi-supervised anomaly detectioncited by 6
  11. 2021CVAD: A generic medical anomaly detector based on Cascade VAEcited by 6
  12. 2021Efficient Anomaly Detection Using Self-Supervised Multi-Cue Taskscited by 6
  13. 2022A Gating Model for Bias Calibration in Generalized Zero-shot Learningcited by 6
  14. 2023Exploiting the Distortion-Semantic Interaction in Fisheye Datacited by 6
  15. 2023Exploring Transformers for Open-world Instance Segmentationcited by 5
  16. 2023Faithfulness Measurable Masked Language Modelscited by 5
  17. 2023Example Forgetting: A Novel Approach to Explain and Interpret Deep Neural Networks in Seismic Interpretationcited by 4
  18. 2024Transitional Uncertainty with Layered Intermediate Predictionscited by 4
  19. 2024Ctrl-GenAug: Controllable Generative Augmentation for Medical Sequence Classificationcited by 4
  20. 2021Anomaly Detection for Solder Joints Using β-VAEcited by 3
  21. 2022Outlier-Robust Group Inference via Gradient Space Clusteringcited by 3
  22. 2023A Survey of Historical Learning: Learning Models with Learning Historycited by 3
  23. 2024Data Valuation with Gradient Similaritycited by 3
  24. 2024VOICE: Variance of Induced Contrastive Explanations to quantify Uncertainty in Neural Network Interpretabilitycited by 3
  25. 2021SLA2P: Self-supervised Anomaly Detection with Adversarial Perturbationcited by 2
  26. 2022R2-AD2: Detecting Anomalies by Analysing the Raw Gradientcited by 2
  27. 2023Stochastic Surprisal: An inferential measurement of Free Energy in Neural Networkscited by 2
  28. 2023Intelligent Anomaly Detection for Lane Rendering Using Transformer with Self-Supervised Pre-Training and Customized Fine-Tuningcited by 2
  29. 2024Approximations to the Fisher Information Metric of Deep Generative Models for Out-Of-Distribution Detectioncited by 2
  30. 2024Counterfactual Gradients-based Quantification of Prediction Trust in Neural Networkscited by 2
  31. 2025GradPCA: Leveraging NTK Alignment for Reliable Out-of-Distribution Detectioncited by 2
  32. 2022Explainable Machine Learning for Hydrocarbon Prospect Riskingcited by 1
  33. 2023Prior Knowledge Guided Network for Video Anomaly Detectioncited by 1
  34. 2024AnomalousPatchCore: Exploring the Use of Anomalous Samples in Industrial Anomaly Detectioncited by 1
  35. 2024Targeting Negative Flips in Active Learning using Validation Setscited by 1
  36. 2022Constraining Representations Yields Models That Know What They Don't Knowcited by 0
  37. 2023RSC-VAE: Recoding Semantic Consistency Based VAE for One-Class Novelty Detectioncited by 0
  38. 2024Effective Data Selection for Seismic Interpretation through Disagreementcited by 0

50 most cited of 54 All 54 (JSON)