Latent Space Autoregression for Novelty Detection

By Davide Abati, Angelo Porrello, Simone Calderara, Rita Cucchiara. Computer Vision and Pattern Recognition, 2018

489 citations recorded by the source corpus.

Novelty detection is commonly referred as the discrimination of observations that do not conform to a learned model of regularity. Despite its importance in different application settings, designing a novelty detector is utterly complex due to the unpredictable nature of novelties and its inaccessibility during the training procedure, factors which expose the unsupervised nature of the problem. In our proposal, we design a general unsupervised framework where we equip a deep autoencoder with a parametric density estimator that learns the probability distribution underlying the latent representations with an autoregressive procedure. We show that a maximum likelihood objective, optimized in conjunction with the reconstruction of normal samples, effectively acts as a regularizer for the task at hand, by minimizing the differential entropy of the distribution spanned by latent vectors. In addition to providing a very general formulation, extensive experiments of our model on publicly available datasets deliver on-par or superior performances if compared to state-of-the-art methods in one-class and in video anomaly detection settings. Differently from our competitors, we remark that our proposal does not make any assumption about the nature of the novelties, making our work easily applicable to disparate contexts.

Subfields: Artificial Intelligence, Computer Networks and Communications, Epidemiology

View Paper: https://www.semanticscholar.org/paper/5f1e7c3c81d6a9e716eab660bb7536ecb204ef7d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Networks and Communications

In Epidemiology

Citations in the corpus

Student-Teacher Feature Pyramid Matching for Anomaly Detection (2021, methodology): “…advance anomaly detection by means of pixel-wise comparison of image patches and their reconstructions [6, 34, 35] or per-pixel estimation of probability density on entire images [1, 37], among which Auto-encoders, Generative Adversarial Networks (GANs), and their variants are dominating models.”
SimpleNet: A Simple Network for Image Anomaly Detection and Localization (2023, methodology): “The baselines include VAE [2], LSA [1], DSVDD [25], OCGAN [19], HRN [15], AnoGAN [27], DAAD [14], MKD [26], DisAug CLR [28], IGD [5] and RevDist [7].”
HRN: A Holistic Approach to One Class Learning (2020, methodology): “class of data using various auto-encoders [1, 14, 65, 71, 83, 69] or GAN [27] style adversarial training [74, 72, 15, 64].”
OpenOOD: Benchmarking Generalized Out-of-Distribution Detection (2022, methodology): “Density-based methods model the in-distribution with probabilistic models, which also achieve good performance and are easier for theoretical analysis [15, 16, 17].”
Multiresolution Knowledge Distillation for Anomaly Detection (2020, background): “To better learn these normal latent features, LSA [1] trains an autoregressive model at its latent space and OC-GAN [31] attempts to force abnormal inputs to be reconstructed as normal ones.”
Learning Deep Features for One-Class Classification (2018, background): “Despite its importance, contemporary one-class classification schemes trained solely on the given concept have failed to produce promising results in real-world datasets [1], [31], [40] has achieved an Area Under the Curve in the range of 60%–65% for CIFAR10 dataset [21].”
Visual Memorability for Robotic Interestingness via Unsupervised Online Learning (2020, unknown): “he interests when it exists for a long time. Nevertheless, the recent approaches of interestingness detection [17,21], as well as saliency detection [41], anomaly detection [42,27], novelty detection [2], and meaningfulness detection [18] algorithms cannot achieve this online updates scheme. To this”
Deep Learning for Anomaly Detection (2020, background): “This research line is popular in video anomaly detection [1, 86, 167].”
PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localization (2020, unknown): “[16], [18], intermediate activations [19] or a discriminator [17], [20] can help to better recognize anomalous images.”
A Unifying Review of Deep and Shallow Anomaly Detection (2020, background): “Currently, the standard evaluation practices in deep AD [130], [135], [137], [141], [144], [148], [152]–[155], [233], [531], OOD detection [241], [277], [490], [494]–[496], [532], [533], and open set recognition [240], [497]–[500] still extensively repurpose classification data sets by deeming some ”
VT-ADL: A Vision Transformer Network for Image Anomaly Detection and Localization (2021, methodology): “COMPARATIVE RESULTS ARE TAKEN FROM [19], [14]”
SSD: A Unified Framework for Self-Supervised Outlier Detection (2021, methodology): “…based detection using Auto-encoders (Hawkins et al., 2002; Mirsky et al., 2018; Schreyer et al., 2017) or Variational auto-encoders (Abati et al., 2019; An & Cho, 2015) 2) Classiﬁcation based, such as Deep-SVDD (Ruff et al., 2018; El-Yaniv & Wiener, 2010; Geifman & El-Yaniv, 2017)…”
Uninformed Students: Student-Teacher Anomaly Detection With Discriminative Latent Embeddings (2019, methodology): “Deep learning based methods for the segmentation of anomalies strongly focus on generative models such as autoencoders [1, 8] or GANs [32].”
FastFlow: Unsupervised Anomaly Detection and Localization via 2D Normalizing Flows (2021, unknown): “2016), LSA (Abati et al. 2019), AnoGAN (Schlegl et al.”
Anomaly Detection via Reverse Distillation from One-Class Embedding (2022, methodology): “The baselines in this experiment include LSA [1], OCGAN [27], HRN [17], DAAD [16] and MKD [33].”
Generalized Out-of-Distribution Detection: A Survey (2021, methodology): “2 can be directly adapted to OOD detection by unifying the ID data as a whole [87], [88], [89], [90], [121].”