Most data collected in real-world scenarios is contaminated, adversely affecting the performance of existing anomaly detection methods that typically rely on clean training data to learn normal patterns. While some approaches have made progress in addressing data contamination, they still face the following challenges: 1) Most methods use pixel-level constraints between the contaminated and reconstructed data as the training loss, and the contaminated data is preferentially learned due to the larger error; 2) Current anomaly detection methods based on frequency-domain do not fully utilize the complex spectral characteristics nor the cross-domain consistency of time series data, which limits the further improvement of the anomaly detection performance. This paper proposes a filter-enhanced reconstruction network with domain-adversarial training for multivariate time series anomaly detection (FERDA), which enables the model to extract normal patterns and high-level semantic information more accurately by training with the error of filtered and reconstructed data. FERDA designs a dual-domain contamination filtering module consisting of the multi-head memory module and the complex linear network. The module is trained to learn normal patterns and cross-domain information through adversarial learning of intra-domain semantic consistency and inter-domain distinguishability. Moreover, a two-stage contamination suppression strategy is proposed. The contamination degree of data is quantified by the difference between pre- and post-filtering, and the impact of contamination is further mitigated by two-stage training with different losses. The proposed method outperforms 23 state-of-the-art baselines in experiments on five benchmark datasets across different domains.

Multivariate time series (MTS), whose patterns change dynamically, often have complex temporal and dimensional dependence. Most existing reconstruction-based MTS anomaly detection methods only learn the point-wise information while ignoring the overall trend of time series, resulting in their incompetence in extracting high-level semantic information. Although a few contrastive learning-based approaches have been proposed recently to solve this problem, they forcibly increase the difference between the features of normal data, leading to the loss of useful information. This paper proposes an adversarial contrastive autoencoder (ACAE) for MTS anomaly detection. ACAE conducts feature combination and decomposition as the contrastive learning proxy task, which introduces adversarial training to learn the transformation-invariant representation of data, achieving a robust representation of MTS. Firstly, ACAE constructs positive and negative sample pairs through the multi-scale timestamp mask and random sampling. Secondly, the features of the original samples are combined with those of the positive and negative samples to generate the positive and negative composite features. Finally, ACAE trains the encoder and discriminator to decompose the negative composite features cooperatively to decrease the similarity between the features of negative pairs. In contrast, it adversarially decomposes the positive composite features to increase the similarity between the features of positive pairs. Experimental results show that ACAE outperforms 14 state-of-the-art baselines on five real-world datasets from different fields.

Nowadays, the massive industrial data has effectively improved the performance of the data-driven deep learning Remaining Useful Life (RUL) prediction method. However, there are still problems of assigning fixed weights to features and only coarse-grained consideration at the sequence level. This paper proposes a Transformer-based end-to-end feature-level mask self-supervised learning method for RUL prediction. First, by proposing a fine-grained feature-level mask self-supervised learning method, the data at different time points under all features in a time window is sent to two parallel learning streams with and without random masks. The model can learn more fine-grained degradation information by comparing the information extracted by the two parallel streams. Instead of assigning fixed weights to different features, the abstract information extracted through the above process is invariable correlations between features, which has a good generalization to various situations under different working conditions. Then, the extracted information is encoded and decoded again using an asymmetric structure, and a fully connected network is used to build a mapping between the extracted information and the RUL. We conduct experiments on the public C-MAPSS datasets and show that the proposed method outperforms the other methods, and its advantages are more obvious in complex multi-working conditions.