AI helps you reading Science
AI generates interpretation videos
AI extracts and analyses the key points of the paper to generate videos automatically
AI parses the academic lineage of this thesis
AI extracts a summary of this paper
We plan to investigate more complex architectures to further improve the accuracy of the proposed HRN method, and exploit the one-class method for positive and unlabeled learning and open-world learning
HRN: A Holistic Approach to One Class Learning
NIPS 2020, (2020)
Existing neural network based one-class learning methods mainly use various forms of auto-encoders or GAN style adversarial training to learn a latent representation of the given one class of data. This paper proposes an entirely different approach based on a novel regularization, called holistic regularization (or H-regularization), whic...More
PPT (Upload PPT)
- One-class learning or classification has many applications. For example, in information retrieval, one has a set of documents of interest and wants to identify more such documents .
- In most applications, deciding whether a data instance belongs to the given training class or is an anomaly can be subjective and a threshold is often used based on the application.
- Like most existing papers [68, 64, 8, 82], this work is interested in a score function instead, and ignores the above binary decision problem
- In this case, the commonly used evaluation metric is AUC (Area Under the ROC curve).
- The authors present the proposed one-class learning method HRN, which uses the above learning paradigm, but employs a novel loss function called one-class loss with an accompanied instance-level data normalization method
- One-class learning or classification has many applications
- Deep learning models have been proposed for the same purpose [68, 8], which mainly learn a good latent representation of the given
- Each row in the table gives the average results over 5 runs of all compared systems using one class as the training data for a dataset
- Existing approaches to one-class learning using deep learning are mainly based on GAN and autoencoders to learn a latent representation of the given class
- This paper proposed an entirely different approach called HRN, which uses a new one-class loss function with a novel regularization method
- We plan to investigate more complex architectures to further improve the accuracy of the proposed HRN method, and exploit the one-class method for positive and unlabeled (PU) learning and open-world learning
- Method OCSVM iForest DAGMM
TQM HRN KDDCUP99
Precision Recall F1
Thyroid Precision Recall
Arrhythmia Precision Recall
Non-Image Datasets: Following the latest baseline TQM , the authors use precision, recall and F1 score as the evaluation measures, and apply the same TQM’s thresholding method in HRN, making the precision, recall and F1 scores the same.
- When our 2-norm instance normalization was added (NLL+Hreg+2N_Inst_Norm), the results were improved even further.
- The instance normalization method in  (NLL+Hreg+Inst_Norm) did fairly well, but it is significantly poorer than HRN (NLL+Hreg+2N_Inst_Norm).
- This is because that the method in  does not solve the problem identified in Sec. 3.2.
- When each new class comes, it is incrementally learned using the one-class loss (Sec. 3.1)
- Results and Discussion
Image Datasets: The authors first report the results on the three image datasets and the three non-image datasets.
- Each row in the table gives the average results over 5 runs of all compared systems using one class as the training data for a dataset.
- Its AUC is only 52.10 while the HRN’s AUC is 71.32
- This dataset was not used in the experimental evaluation of the TQM paper.
- Its average AUC on the CIFAR-10 dataset is much better than other baselines, it is still markedly lower than HRN.
- Existing approaches to one-class learning using deep learning are mainly based on GAN and autoencoders to learn a latent representation of the given class.
- This paper proposed an entirely different approach called HRN, which uses a new one-class loss function with a novel regularization method.
- The architecture of HRN is very simple.
- The authors plan to investigate more complex architectures to further improve the accuracy of the proposed HRN method, and exploit the one-class method for PU learning and open-world learning
- Table1: Average AUCs in % over 5 runs per method on the three image datasets
- Table2: Average precision, recall, and F1 score on the three non-image datasets over 5 runs
- Table3: Average AUCs in % of different components of HRN on the image datasets. Hreg: H-regularization; 2N_Inst_Norm: our 2-norm instance normalization; Inst_Norm[<a class="ref-link" id="c79" href="#r79">79</a>]: instance normalization in [<a class="ref-link" id="c79" href="#r79">79</a>]; SquareLoss: Square Loss
- Table4: Average AUCs in % on CIFAR-10: Pre-training using ImageNet without overlapping classes
- Table5: Average AUCs in % with different training data noise ratio for HRN
- Table6: Average AUCs in % per method on GTSRB stop sign adversarial attacks
- Table7: Continual learning accuracy results for 1 class per task of HCL and the baselines
- Much of the existing work on anomaly, outlier or novelty detection can be regarded as some form of one-class learning from a class of normal data. Early work in statistics  was mainly based on probabilistic modeling of the distribution of the normal data and regard data points with low probabilities in the distribution as anomalies [4, 87, 22, 86]. In general, anomaly detection algorithms can be classified into the following categories: distance based methods [42, 3, 28, 31, 60], density based methods [38, 56, 6], mixture models [3, 43], one-class classification based methods [75, 78, 39], deep learning based representation learning using auto-encoders [10, 89, 69, 7, 94] and adversarial learning [74, 15, 64, 20], ensemble methods [53, 10], graphs and random walks [58, 31], transfer learning [45, 2], and multi-task learning . Several surveys have also been published [9, 66, 7, 61].
About one-class learning, one-class SVM (OCSVM)  was perhaps the earliest method, which uses the kernel SVM to separate the data from the origin. It essentially treats the origin as the only negative data point. Another earlier method based on kernel SVM is the Support Vector Data Description (SVDD) , which tries to find a hypersphere to enclose the given class of data.  learns features using deep learning and then applies OCSVM or SVDD to build the one-class model.
- Acknowledgments and Disclosure of Funding This work was partially supported by the National Key Research and Development Program of China under grant 2018AAA0100205
Study subjects and analysis
Table 3 gives the ablation results of MNIST, fMNIST and CIFAR-10. We see that using only NLL, the model performed very poorly on all three datasets. Including H-regularization (NLL+Hreg) improves the performance drastically
image classification datasets: 3
To our knowledge, both H-regularization and the normalization method have not been reported in the literature. Empirical evaluation using three image classification datasets widely used in evaluating one-class learners and three traditional benchmark anomaly detection datasets demonstrates the effectiveness of HRN. It outperforms eleven state-of-the-art baselines considerably
benchmark datasets: 6
4 Empirical Evaluation. We empirically evaluate the proposed algorithm HRN using six benchmark datasets and eleven state-of-the-art baselines. Following existing papers, no pre-trained feature extractors were used in the main evaluation
benchmark image classification datasets: 3
Datasets. We use three benchmark image classification datasets and three benchmark traditional nonimage anomaly detection datasets that have been used in many previous papers. (1) MNIST 6 is a handwritten digit classification dataset of 10 digits, i.e., 10 classes
image datasets: 3
(3) CIFAR-10 8 is also an image classification dataset consisting of 60,000 32x32 color images of 10 classes with the splitting of 50,000 for training and 10,000 for testing. For each of these three image datasets, we use the training data of each class C in the dataset in turn as the one class data to build a model and then test the model using the full test set of all classes. The rest of the classes except C are anomalies
non-image datasets: 3
The rest of the classes except C are anomalies. The three non-image datasets are: (4) KDDCUP99 9 consists of 450000 training instances and 44021 test instances of two classes. The majority class (80% of the data) is regarded as the one class used in learning
We followed the TQM approach and used grid search. However, we used only the MNIST data to search for hyper-parameter values and then applied the values to all 5 datasets. Grid search uses the following tuning ranges: for λ, from 0 to 1 with step 0.05 and for n, from 1 to 20 with step 1
image datasets: 3
4.3 Results and Discussion. Image Datasets: We first report the results on the three image datasets and then the three non-image datasets. The main results on the three image datasets are given in Table 1.13 AUC (Area Under the ROC curve) is the evaluation metric, which is also used in most one-class and other anomaly detection algorithms [7, 64, 68]
image datasets: 3
Image Datasets: We first report the results on the three image datasets and then the three non-image datasets. The main results on the three image datasets are given in Table 1.13 AUC (Area Under the ROC curve) is the evaluation metric, which is also used in most one-class and other anomaly detection algorithms [7, 64, 68]. Each row in the table gives the average results over 5 runs of all compared systems using one class (column 1) as the training data for a dataset
- Davide Abati, Angelo Porrello, Simone Calderara, and Rita Cucchiara. Latent space autoregression for novelty detection. In CVPR, 2019.
- J. T. Andrews, T. Tanay, E. J. Morton, and L. D. Griffin. Transfer representation-learning for anomaly detection. In Anomaly Detection Workshop in ICML, 2016.
- Fabrizio Angiulli and Clara Pizzuti. Fast outlier detection in high dimensional spaces. In Principles of Data Mining and Knowledge Discovery, 2002.
- Vic Barnett and Toby Lewis, editors. Outliers in statistical data. Wiley, 1974.
- Wieland Brendel, Jonas Rauber, and Matthias Bethge. Decision-based adversarial attacks: Reliable attacks against black-box machine learning models. arXiv preprint arXiv:1712.04248, 2017.
- Markus M. Breunig, Hans-Peter Kriegel, Raymond T. Ng, and Jörg Sander. Lof: Identifying density-based local outliers. In SIGMOD, 2000.
- Raghavendra Chalapathy and Sanjay Chawla. Deep learning for anomaly detection: A survey. In arXiv:1901.03407v2 [cs.LG], 2019.
- Raghavendra Chalapathy, Aditya Krishna Menon, and Sanjay Chawla. Anomaly detection using one-class neural networks. In arXiv:1802.06360 [cs.LG], 2018.
- Varun Chandola, Arindam Banerjee, and Vipin Kumar. Anomaly detection: A survey. In ACM Computing Surveys, 2009.
- Jinghui Chen, Saket Sathe, Charu C. Aggarwal,, and Deepak S. Turaga. Outlier detection with autoencoder ensembles. In SDM, 2017.
- Zhiyuan Chen and Bing Liu. Lifelong machine learning. Synthesis Lectures on Artificial Intelligence and Machine Learning, 12(3):1–207, 2018.
- Gregory Cohen, Saeed Afshar, Jonathan Tapson, and Andre van Schaik. Emnist: an extension of mnist to handwritten letters. http://arxiv.org/abs/1702.05373, 2017.
- Yang Cong, Junsong Yuan, and Ji Liu. Sparse reconstruction cost for abnormal event detection. In CVPR, 2011.
- Sanjoy Dasgupta, Timothy C. Sheehan, Charles F. Stevens, and Saket Navlakha. A neural data structure for novelty detection. In Proceedings of the National Academy of Sciences, 2018.
- Lucas Deecke, Robert Vandermeulen, Lukas Ruff, Stephan Mandt, and Marius Klof. Image anomaly detection with generative adversarial networks. In ECML/PKDD, 2018.
- Lucas Deecke, Robert Vandermeulen, Lukas Ruff, Stephan Mandt, and Marius Kloft. Anomaly detection with generative adversarial networks. In ECML PKDD 2018, 2019.
- Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. Bert: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-2019), 2019.
- Terrance DeVries and Graham W Taylor. Learning confidence for out-of-distribution detection in neural networks. arXiv preprint arXiv:1802.04865, 2018.
- Yihe Dong, Samuel B. Hopkins, and Jerry Li. Quantum entropy scoring for fast robust mean estimation and improved outlier detection. In NeurIPS, 2019.
- Sarah M Erfani, Mahsa Baktashmotlagh, Masud Moshtaghi, Vinh Nguyen, Christopher Leckie, James Bailey, and Kotagiri Ramamohanarao. From shared subspaces to shared landmarks: A robust multi-source classification approach. In AAAI, 2017.
- Sarah M Erfani, Sutharshan Rajasegarar, Shanika Karunasekera, and Christopher Leckie. Highdimensional and largescale anomaly detection using a linear one-class svm with deep learning. Pattern Recognition, 58:121–134, 2016.
- Eleazar Eskin. Anomaly detection over noisy data using learned probability distributions. In ICML, 2000.
- Geli Fei and Bing Liu. Breaking the closed world assumption in text classification. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 506–514, 2016.
- Geli Fei, Shuai Wang, and Bing Liu. Learning cumulatively to become more knowledgeable. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 1565–1574, 2016.
- Izhak Golan and Ran El-Yaniv. Deep anomaly detection using geometric transformations. In NeurIPS, 2018.
- Markus Goldstein and Seiichi Uchida. A comparative evaluation of unsupervised anomaly detection algorithms for multivariate data. PloS one, 11(4):e0152173, 2016.
- Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. In NIPS-2014, 2014.
- Xiaoyi Gu, Leman Akoglu, and Alessandro Rinaldo. Statistical analysis of nearest neighbor methods for anomaly detection. In NeurIPS, 2019.
- Ishaan Gulrajani, Faruk Ahmed, Martin Arjovsky, Vincent Dumoulin, and Aaron C Courville. Improved training of wasserstein gans. In Advances in neural information processing systems, pages 5767–5777, 2017.
- Raia Hadsell, Sumit Chopra, and Yann LeCun. Dimensionality reduction by learning an invariant mapping. In CVPR, 2006.
- Ville Hautamaki, Ismo Karkkainen, and Pasi Franti. Outlier detection using k-nearest neighbour graph. In ICPR, 2004.
- Kaiming He, Ross Girshick, and Piotr Dollar. Rethinking imagenet pre-training. arXiv:1811.08883v1 [cs.CV], 2018.
- D. Hendrycks and K. Gimpel. A baseline for detecting misclassified and out-of-distribution examples in neural networks. In ICLR, 2017.
- Wenpeng Hu, Zhou Lin, Bing Liu, Chongyang Tao, Zhengwei Tao, Jinwen Ma, Dongyan Zhao, and Rui Yan. Overcoming catastrophic forgetting for continual learning via model adaptation. In ICLR, 2019.
- T. Ide, D. T. Phan, and J. Kalagnanam. Multi-task multi-modal models for collective anomaly detection. In ICDM, 2017.
- Vilen Jumutc and Johan AK Suykens. Multi-class supervised novelty detection. IEEE transactions on pattern analysis and machine intelligence, 36(12):2510–2523, 2014.
- Shehroz S. Khan and Michael G. Madden. One-class classification: taxonomy of study and review of techniques. The Knowledge Engineering Review, 29(3):345–374, 2014.
- JooSeuk Kim and Clayton D. Scott. Robust kernel density estimation. Journal of Machine Learning Research, 13(1):2529–2565, 2012.
- Sangwook Kim, Yonghwa Choi, and Minho Lee. Deep learning with support vector data description. Neurocomputing, 165:111–117, 2015.
- Diederik P Kingma and Max Welling. Auto-encoding variational bayes. In ICLR, 2014.
- James Kirkpatrick, Razvan Pascanu, Neil Rabinowitz, Joel Veness, Guillaume Desjardins, Andrei A Rusu, Kieran Milan, John Quan, Tiago Ramalho, Agnieszka Grabska-Barwinska, and Others. Overcoming catastrophic forgetting in neural networks. volume 114, pages 3521–3526. National Acad Sciences, 2017.
- Edwin M Knorr, Raymond T Ng, and Vladimir Tucakov. Distance-based outliers: algorithms and applications. The VLDB Journal, 8(3-4), 2000.
- Hans-Peter Kriegel, Matthias Schubert, and Arthur Zimek. Angle-based outlier detection in high-dimensional data. In KDD, 2008.
- Alex Krizhevsky and Geoffrey Hinton. Learning multiple layers of features from tiny images. Technical Report TR-2009, University of Toronto, Toronto., 2009.
- Atsutoshi Kumagai, Tomoharu Iwata, and Yasuhiro Fujiwara. Transfer anomaly detection by inferring latent domain representations. In NeurIPS, 2019.
- Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, and Radu Soricut. Albert: A lite bert for self-supervised learning of language representations. arXiv:1909.11942v1 [cs.CL], 2019.
- Yann LeCun, Corinna Cortes, and Christopher JC Burges. The mnist database of handwritten digits. http://yann.lecun.com/exdb/mnist/, 1998.
- Sang-Woo Lee, Jin-Hwa Kim, Jaehyun Jun, Jung-Woo Ha, and Byoung-Tak Zhang. Overcoming catastrophic forgetting by incremental moment matching. In NIPS, pages 4655–4665, 2017.
- W. Li, V. Mahadevan, and N. Vasconcelos. Anomaly detection and localization in crowded scenes. IEEE Transactions on Pattern Analysis and Machine Intelligence, 36(1):18–22, 2014.
- Zhizhong Li and Derek Hoiem. Learning without forgetting. PAMI, 40(12):2935–2947, 2017.
- Shiyu Liang, Yixuan Li, and R. Srikant. Enhancing the reliability of out-of-distribution image detection in neural networks. In ICLR, 2018.
- Bing Liu, Yang Dai, Xiaoli Li, Wee Sun Lee, and Philip S. Yu. Building text classifiers using positive and unlabeled examples. In ICDM, 2003.
- Fei Tony Liu, Kai Ming Ting, and Zhi-Hua Zhou. Isolation forest. In ICDM, pages 413–422, 2008.
- Yinhan Liu, Myle Ott, Jingfei Du Naman Goyal, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692, 2019.
- Larry Manevitz and Malik Yousef. One-class document classification via neural networks. Neurocomputing, 70(7-9):1466–1481, 2007.
- Emaad Manzoor, Hemank Lamba, and Leman Akoglu. xstream: Outlier dete‘x’ion in featureevolving data streams. In KDD-2018, 2018.
- Michael McCloskey and Neal J Cohen. Catastrophic interference in connectionist networks: The sequential learning problem. In Psychology of learning and motivation. 1989.
- HDK Moonesinghe and Pang-Ning Tan. Outrank: a graph-based outlier detection framework using random walk. International Journal on Artificial Intelligence Tools, 17(1):19–36, 2008.
- Duc Tam Nguyen, Zhongyu Lou, and Michael Klarand2 Thomas Brox. Anomaly detection with multiple-hypotheses predictions. In ICML, 2019.
- Guansong Pang, Longbing Cao, Ling Chen, and Huan Liu. Learning representations of ultrahighdimensional data for random distance-based outlier detection. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 2041–2050, 2018.
- Guansong Pang, Chunhua Shen, Longbing Cao, and Anton van den Hengel. Deep learning for anomaly detection: A review. arXiv preprint arXiv:2007.02500, 2020.
- Guansong Pang, Chunhua Shen, and Anton van den Hengel. Deep anomaly detection with deviation networks. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 353–362, 2019.
- German Ignacio Parisi, Ronald Kemker, Jose L. Part, Christopher Kanan, and Stefan Wermter. Continual lifelong learning with neural networks: A review. Neural Networks, 2019.
- Pramuditha Perera, Ramesh Nallapati, and Bing Xiang. Ocgan: One-class novelty detection using gans with constrained latent representations. In CVPR, 2019.
- Stanislav Pidhorskyi, Ranya Almohsen, and Gianfranco Doretto. Generative probabilistic novelty detection with adversarial autoencoders. In NeurIPS-2018, 2018.
- Marco A.F. Pimentel, David A. Clifton, Lei Clifton, and Lionel Tarassenko. A review of novelty detection. Signal Processing, 99:215–249, 2014.
- Jathushan Rajasegaran, Munawar Hayat, Salman H Khan, Fahad Shahbaz Khan, and Ling Shao. Random path selection for continual learning. In H. Wallach, H. Larochelle, A. Beygelzimer, F. dAlché Buc, E. Fox, and R. Garnett, editors, NIPS, pages 12648–12658. Curran Associates, Inc., 2019.
- Lukas Ruff, Robert Vandermeulen, Nico Goernitz, Lucas Deecke, Shoaib Ahmed Siddiqui, Alexander Binder, Emmanuel Müller, and Marius Kloft. Deep one-class classification. In ICML, 2018.
- M Sabokrou, M Fathy, and M Hoseini. Video anomaly detection and localisation based on the sparsity and reconstruction error of auto-encoder. Electronics Letters, 52(13):1122–1124, 2016.
- M. Sabokrou, M. Fayyaz, M. Fathy, and R. Klette. Deep-cascade: Cascading 3d deep neural networks for fast anomaly detection and localization in crowded scenes. IEEE Transactions on Image Processing, 26(4):1992–2004, 2017.
- M. Sabokrou, M. Khalooei, M. Fathy, and E. Adeli. One-class classifier for novelty detection. In CVPR, 2018.
- Mohammad Sabokrou, Mohammad Khalooei, Mahmood Fathy, and Ehsan Adeli. Adversarially learned one-class classifier for novelty detection. In CVPR, 2018.
- Patrick Schlachter, Yiwen Liao, and Bin Yang. Deep one-class classification using intra-class splitting. arXiv:1902.01194v4 [cs.LG], 2019.
- Thomas Schlegl, Philipp Seebock, Sebastian M. Waldstein, Ursula Schmidt-Erfurth, and Georg Langs. Unsupervised anomaly detection with generative adversarial networks to guide marker discovery. In IPMI, 2017.
- Bernhard Schölkopf, John C. Platt, John C. Shawe-Taylor, Alex J. Smola, and Robert C. Williamson. Estimating the support of a high-dimensional distribution. In Neural Computation, 2001.
- Lei Shu, Hu Xu, and Bing Liu. Unseen class discovery in open-world classification. arXiv:1801.05609, 2018.
- Linda Studer, Michele Alberti, Vinaychandran Pondenkandath, Pinar Goktepe, Thomas Kolonko, Andreas Fischer, Marcus Liwicki, and Rolf Ingold. A comprehensive study of imagenet pretraining for historical document image analysis. arXiv:1905.09113 [cs.CV], 2019.
- David M.J. Tax and Robert P.W. Duin. Support vector data description. Machine Learning, 54(1):45–66, 2004.
- Dmitry Ulyanov and Andrea Vedaldi. Instance normalization: The missing ingredient for fast stylization. https://arxiv.org/pdf/1607.08022.pdf, 2017.
- Pascal Vincent, Hugo Larochelle, Yoshua Bengio, and Pierre-Antoine Manzagol. Extracting and composing robust features with denoising autoencoders. In ICML, 2008.
- Jingjing Wang, Sun Sun, and Yaoliang Yu. Multivariate triangular quantile maps for novelty detection. In NeurIPS, 2019.
- Siqi Wang, Yijie Zeng, Xinwang Liu, En Zhu, Jianping Yin, Chuanfu Xu, and Marius Kloft. Effective end-to-end unsupervised outlier detection via inlier priority of discriminative network. In NeurIPS, 2019.
- Y. Xia, X. Cao, F. Wen, G. Hua, and J. Sun. Learning discriminative reconstructions for unsupervised outlier removal. In CVPR, 2015.
- Han Xiao, Kashif Rasul, and Roland Vollgraf. Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms. rXiv preprint arXiv:1708.07747, 2017.
- Hu Xu, Bing Liu, Lei Shu, and P Yu. Open-world learning and application to product classification. In The World Wide Web Conference, pages 3413–3419, 2019.
- Huan Xu, Constantine Caramanis, and Sujay Sanghavi. Robust pca via outlier pursuit. In NIPS, 2010.
- Kenji Yamanishi, Jun-Ichi Takeuchi, Graham Williams, and Peter Milne. On-line unsupervised outlier detection using finite mixtures with discounting learning algorithms. Data Mining and Knowledge Discovery, 8(3):275–300, 2004.
- Zhilin Yang, Zihang Dai, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, and Quoc V Le. Xlnet: Generalized autoregressive pretraining for language understanding. arXiv preprint arXiv:1906.08237, 2019.
- Chong You, Daniel P Robinson, and René Vidal. Provable self-representation based outlier detection in a union of subspaces. CVPR, 2017.
- Sergey Zagoruyko and Nikos Komodakis. Wide residual networks. arXiv preprint arXiv:1605.07146, 2016.
- Guanxiong Zeng, Yang Chen, Bo Cui, and Shan Yu. Continuous learning of context-dependent processing in neural networks. Nature Machine Intelligence, 2019.
- Shuangfei Zhai, Yu Cheng, Weining Lu, and Zhongfei Zhang. Deep structured energy based models for anomaly detection. In ICML, 2016.
- Panpan Zheng, Shuhan Yuan, Xintao Wu, Jun Li, and Aidong Lu. One-class adversarial nets for fraud detection. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, pages 1286–1293, 2019.
- Chong Zhou and Randy C Paffenroth. Anomaly detection with robust deep autoencoders. In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 665–674, 2017.
- Bo Zong, Qi Song, Martin Renqiang Min, Wei Cheng, Cristian Lumezanu, Daeki Cho, and Haifeng Chen. Deep autoencoding gaussian mixture model for unsupervised anomaly detection. In ICLR, 2018.