Journal Article10.1109/tii.2023.3345461
Can Untrained Neural Networks Detect Anomalies?
Seunghyoung Ryu,Yonggyun Yu,Hogeon Seo +2 more
1
TL;DR: It is demonstrated that UNNs can achieve competitive AD performance without training, which also underscores the importance of training to ensure higher performance beyond the untrained baseline, making it a compelling alternative for various applications.
read more
Abstract: Anomaly detection (AD) plays a crucial role in identifying unusual data patterns indicative of potential issues or opportunities. Recent data-driven AD models require extensive training for satisfactory performance. This study explores the potential of untrained neural networks (UNNs) for AD tasks. UNNs are used for nonlinear random projection. The anomaly scores are derived from the randomly mapped features using the Mahalanobis distance. We conducted a series of experiments on 12 tabular and two image datasets, comparing the performance of UNNs with 12 established AD models, including state-of-the-art deep learning approaches. Our results demonstrate that UNNs can achieve competitive AD performance without training, which also underscores the importance of training to ensure higher performance beyond the untrained baseline. In addition, the proposed approach offers advantages in terms of time, computational costs, and accessibility, making it a compelling alternative for various applications.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Change Representation and Extraction in Stripes: Rethinking Unsupervised Hyperspectral Image Change Detection With an Untrained Network
Bin Yang,Yin Mao,Licheng Liu,Leyuan Fang,Xinxin Liu +4 more
TL;DR: This study proposes StripeCD, an unsupervised hyperspectral image change detection method using an untrained network, which represents and models changes in stripes, outperforming state-of-the-art approaches on three datasets and indicating potential for further untrained network applications.
5
References
•Proceedings Article
Deep One-Class Classification
Lukas Ruff,Robert A. Vandermeulen,Nico Goernitz,Lucas Deecke,Shoaib Ahmed Siddiqui,Alexander Binder,Emmanuel Müller,Marius Kloft +7 more
- 03 Jul 2018
TL;DR: This paper introduces a new anomaly detection method—Deep Support Vector Data Description—, which is trained on an anomaly detection based objective and shows the effectiveness of the method on MNIST and CIFAR-10 image benchmark datasets as well as on the detection of adversarial examples of GTSRB stop signs.
Database-friendly random projections: Johnson-Lindenstrauss with binary coins
TL;DR: Two constructions of k-dimensional Euclidean embeddings with the property that all elements of the projection matrix belong in {-1, 0, +1 } are given.
1.7K
GANomaly : semi-supervised anomaly detection via adversarial training.
Samet Akcay,Amir Atapour-Abarghouei,Toby P. Breckon +2 more
- 02 Dec 2018
TL;DR: In this paper, a conditional generative adversarial network (GAN) is used for anomaly detection in a one-class, semi-supervised learning paradigm, where an encoder-decoder-encoder sub-network is employed to map the input image to a lower dimension vector, which is then used to reconstruct the generated output image.
Anomaly Detection with Robust Deep Autoencoders
Chong Zhou,Randy Paffenroth +1 more
- 04 Aug 2017
TL;DR: Novel extensions to deep autoencoders are demonstrated which not only maintain a deep autenkocoders' ability to discover high quality, non-linear features but can also eliminate outliers and noise without access to any clean training data.
1.6K
f-AnoGAN: Fast unsupervised anomaly detection with generative adversarial networks.
TL;DR: Fast AnoGAN (f‐AnoGAN), a generative adversarial network (GAN) based unsupervised learning approach capable of identifying anomalous images and image segments, that can serve as imaging biomarker candidates is presented.
1.4K