Time series anomaly detection is typically used to identify data that deviates significantly from normal data, often indicating faults or failures in the underlying system, thus facilitating system stability and safety. Most existing methods focus on detecting anomalies after they occur, while research on predicting future anomalies remains scarce. Before anomalies manifest themselves, there are often subtle precursors exhibiting slight deviations from normal behavior, with varying reaction times and intensities. Next, the setting is often characterized by a lack of labeled data, which complicates model training. To address these challenges, we propose a time series anomaly prediction framework, TAP. It can adapt flexibly to varying reaction times of anomaly precursors across different variables using a period-aware multi-scale module, and it is able to strengthen the distinction between precursors and normal sequences via a dual-branch framework that combines reconstruction and contrastive learning. The contrastive branch employs a controlled generation strategy within the multi-scale patching to produce diverse hard negative samples for precursor identification. The reconstruction branch complements this by evaluating fluctuation magnitudes to ensure sensitivity to subtle variations. We report on experiments on eight datasets from diverse domains, finding that TAP is capable of competitive or superior performance compared to baseline methods for both anomaly detection and prediction.

Model quantization is an effective way to compress deep neural networks and accelerate the inference time on edge devices. Existing quantization methods usually require original data for calibration during the compressing process, which may be inaccessible due to privacy issues. A common way is to generate calibration data to mimic the origin data. However, the generators in these methods have the mode collapse problem, making them unable to synthesize diverse data. To solve this problem, we leverage the information from the full-precision model and enhance both inter-class and intra-class diversity for generating better calibration data, by devising a multi-layer features mixer and normalization flow based attention. Besides, novel regulation losses are proposed to make the generator produce diverse data with more patterns from the perspective of activated feature values and for the quantized model to learn better clip ranges adaptive to our diverse calibration data. Extensive experiments show that our method achieves state-of-the-art quantization results for both Transformer and CNN architectures. In addition, we visualize the generated data to verify that our strategies can effectively handle the mode collapse issue. Our codes are available at repo.

Time series anomaly detection plays a vital role in a wide range of applications. Existing methods require training one specific model for each dataset, which exhibits limited generalization capability across different target datasets, hindering anomaly detection performance in various scenarios with scarce training data. Aiming at this problem, we propose constructing a general time series anomaly detection model, which is pre-trained on extensive multi-domain datasets and can subsequently apply to a multitude of downstream scenarios. The significant divergence of time series data across different domains presents two primary challenges in building such a general model: (1) meeting the diverse requirements of appropriate information bottlenecks tailored to different datasets in one unified model, and (2) enabling distinguishment between multiple normal and abnormal patterns, both are crucial for effective anomaly detection in various target scenarios. To tackle these two challenges, we propose a general time series anomaly Detector with Adaptive Bottlenecks and Dual Adversarial Decoders (DADA), which enables flexible selection of bottlenecks based on different data and explicitly enhances clear differentiation between normal and abnormal series. We conduct extensive experiments on nine target datasets from different domains. After pre-training on multi-domain data, DADA, serving as a zero-shot anomaly detector for these datasets, still achieves competitive or even superior results compared to those models tailored to each specific dataset.

With the growing availability of multi-domain time series data, there is an increasing demand for general forecasting models pre-trained on multi-source datasets to support diverse downstream prediction scenarios. Existing time series foundation models primarily focus on scaling up pre-training datasets and model sizes to enhance generalization performance. In this paper, we take a different approach by addressing two critical aspects of general forecasting models: (1) how to derive unified representations from heterogeneous multi-domain time series data, and (2) how to effectively capture domain-specific features to enable adaptive transfer across various downstream scenarios. To address the first aspect, we propose Decomposed Frequency Learning as the pre-training task, which leverages frequency-based masking and reconstruction to decompose coupled semantic information in time series, resulting in unified representations across domains. For the second aspect, we introduce the Time Series Register, which captures domain-specific representations during pre-training and enhances adaptive transferability to downstream tasks. Our model achieves the state-of-the-art forecasting performance on seven real-world benchmarks, demonstrating remarkable few-shot and zero-shot capabilities.

We study the problem of time series anomaly prediction, which is relevant to a range of real-world applications. Existing anomaly prediction methods rely on labeled training data for achieving acceptable accuracy. However, such data may be difficult to obtain; and in real-time deployments, anomalies can occur that were not seen in labeled data, thus making them difficult to predict. We provide a theoretical analysis and propose an Importance-based Generative Contrastive Learning method (IGCL) for unsupervised anomaly prediction. IGCL employs a controlled diffusion module to produce anomaly precursor patterns. Next, ICGL learns contextual representations to extract temporal dependencies from pairs of normal time series and anomaly precursors. IGCL is then able to predict anomalies by identifying anomaly precursors that will evolve into future anomalies. To address challenges caused by potentially complex precursor combinations involving multiple variables, we propose a memory bank with importance scores that stores representative samples adaptively and generates more complex anomaly precursors. Extensive experiments on nine benchmark datasets offer evidence that the proposed method is able to outperform state-of-the-art baselines.

Accurate time series forecasting is crucial for optimizing resource allocation, industrial production, and urban management, particularly with the growth of cyber-physical and IoT systems. However, limited training sample availability in fields like physics and biology poses significant challenges. Existing models struggle to capture long-term dependencies and to model diverse meta-knowledge explicitly in few-shot scenarios. To address these issues, we propose MetaGP, a meta-learning-based Gaussian process latent variable model that uses a Gaussian process kernel function to capture long-term dependencies and to maintain strong correlations in time series. We also introduce Kernel Association Search (KAS) as a novel meta-learning component to explicitly model meta-knowledge, thereby enhancing both interpretability and prediction accuracy. We study MetaGP on simulated and real-world few-shot datasets, showing that it is capable of state-of-the-art prediction accuracy. We also find that MetaGP can capture long-term dependencies and can model meta-knowledge, thereby providing valuable insights into complex time series patterns.

Accurate long-term forecasting from multivariate time series has important real-world applications. However, achieving this so is challenging. Thus, analyses reveal that time series that span long durations often exhibit dynamic and disrupted correlations. State-of-the-art methods employ attention mechanisms to capture dynamic correlations, but they often do not contend well with disrupted correlations, which reduces prediction accuracy. We introduce local and global information concepts and then leverage these in a Memory Guided Transformer, called the Memformer. By integrating patch-wise recurrent graph learning and global attention, the Memformer aims to capture dynamic correlations and take disrupted correlations into account. We also integrate a so-called Alternating Memory Enhancer into the Memformer to capture correlations between local and global information. We report on experiments that offer insight into the effectiveness of the Memformer at capturing dynamic correlations and its robustness to disrupted correlations. The experiments offer evidence that the new method is capable of advancing the state-of-the-art in forecasting accuracy on real-world datasets.

Flink clusters often suffer from hotspot issues where the monitored job delay and CPU usage keep rising and remain high. This necessitates the detection of anomalous time series to pinpoint the hotspot machines. However, the state-of-the-art unsupervised time series anomaly detection (UTAD) methods are ineffective in this scenario. We identify two main reasons for this. First, the hotspot scenario requires us to pay particular attention to Flink-specific anomalies, e.g., slow-rising and high-level anomalies, which the existing methods struggle to address. Second, the state-of-the-art anomaly detection methods often assume that training datasets do not contain anomalies, but the data collected from the running Flink clusters contains noise, which causes these methods to learn anomalous patterns as normal patterns. In this paper, we first conduct experiments to analyze why existing methods fail in the Flink scenario. To tackle these challenges, we propose a cross-contrastive approach to learn the context information for each timestamp to enable Flink-specific anomaly detection. Then, to address noisy anomalies, we incorporate prior knowledge to set an anomaly boundary to prevent the model from learning anomalous patterns. Extensive experiments show that our method not only outperforms existing methods in the Flink scenario but also achieves state-of-the-art results on public benchmark datasets.

Robust multivariate time series forecasting is crucial in many cyberphysical and Internet of Things applications. Existing state-of-the-art robust forecasting models decompose time series into independent functions covering trends and periodicities. However, these independent functions fail to capture correlations among multiple time series, thereby reducing prediction accuracy. Moreover, existing robust forecasting models treat certain abrupt but normal changes, e.g., caused by holidays, as outliers because they occur infrequently and have data distributions that resemble those of outliers. This exacerbates model bias and reduces prediction accuracy. This paper aims to capture correlations across multiple time series and abrupt but normal changes, thereby improving prediction accuracy. We employ weak labels to partition the dataset into source and target domains. Then, we propose the Domain Adversarial Robust Forecaster (DARF). This forecasting model is based on adversarial domain adaptation and includes two novel modules: Correlated Robust Forecaster (CORF) and Domain Critic. Specifically, CORF constitutes an encoder-decoder framework proficient at robust multivariate time series forecasting, and Domain Critic works to reduce data bias. Extensive experiments and discussions show that DARF is capable of state-of-the-art forecasting accuracy.

Multiple time series forecasting plays an essential role in many applications. Solutions based on graph neural network (GNN) that deliver state-of-the-art forecasting performance use the relation graph which can capture historical correlations among time series. However, in real world, it is common that correlations among time series evolve across time, resulting in dynamic relation graph, where the future correlations may be different from those in history. To address this problem, we propose multiple time series forecasting with dynamic graph modeling (MTSF-DG) that is able to learn historical relation graphs and predicting future relation graphs to capture the dynamic correlations. We also propose a causal GNN to extract features from both kinds of relation graphs efficiently. Then we propose a reasoning network to explicitly learn the variant influence from historical timestamps to future timestamps for final forecasting. Extensive experiments on six benchmark datasets show that MTSF-DG consistently outperforms state-of-the-art baselines, and justify our design with dynamic relation graph modeling.

Click-through rate (CTR) prediction plays an important role in search and recommendation, which are the two most prominent scenarios in e-commerce. A number of models have been proposed to predict CTR by mining user behaviors, especially users' interactions with items. But the sparseness of user behaviors is an obstacle to the improvement of CTR prediction. Previous works only focused on one scenario, either search or recommendation. However, on a practical e-commerce platform, search and recommendation share the same set of users and items, which means joint learning of both scenarios may alleviate the sparseness of user behaviors. In this paper, we propose a novel Search and Recommendation Joint Graph (SRJGraph) neural network to jointly learn a better CTR model for both scenarios. A key question of joint learning is how to effectively share information across search and recommendation, in spite of their differences. A notable difference between search and recommendation is that there are explicit queries in search, whereas no query exists in recommendation. We address this difference by constructing a unified graph to share representations of users and items across search and recommendation, as well as represent user-item interactions uniformly. In this graph, users and items are heterogeneous nodes, and search queries are incorporated into the user-item interaction edges as attributes. For recommendation where no query exists, a special attribute is attached on user-item interaction edges. We further propose an intention and upstream-aware aggregator to explore useful information from high-order connections among users and items. We conduct extensive experiments on a large-scale dataset collected from Taobao.com, the largest e-commerce platform in China. Empirical results show that SRJGraph significantly outperforms the state-of-the-art approaches of CTR prediction in both search and recommendation tasks.

Multiplex networks contain multiple types of relations between nodes, where each relation type is modeled as one layer. In the real world, a relation type may only depend on certain attribute features of nodes. Most existing multiplex network embedding methods only focus on preserving consistent information or complementary information from multiplex networks. However, these methods ignore the dependency between node attributes and the topology of each relation. To address the problem, we propose a model called DAME (Disentangled-based Adversarial Network for Multiplex Network Embedding). We utilize generative adversarial learning to preserve the consistent and complementary information between different relation types. Meanwhile, we develop a disentangled graph convolution network (DGCN) based on disentangled learning, enabling DAME to capture the dependency between node attributes and each relation type. We conduct extensive experiments on five realworld datasets. Experimental results indicate the effectiveness of DAME on link prediction and node classification tasks.

Heterogeneous information network (HIN) contains multiple types of entities and relations. Most of existing HIN embedding methods learn the semantic information based on the heterogeneous structures between different entities, which are implicitly assumed to be complete. However, in real world, it is common that some relations are partially observed due to privacy or other reasons, resulting in a sparse network, in which the structure may be incomplete, and the ”unseen” links may also be positive due to the missing relations in data collection. To address this problem, we propose a novel and principled approach: a Multi-View Adversarial Completion Model (MV-ACM). Each relation space is characterized in a single viewpoint, enabling us to use the topological structural information in each view. Based on the multi-view architecture, an adversarial learning process is utilized to learn the reciprocity (i.e., complementary information) between different relations: In the generator, MV-ACM generates the complementary views by computing the similarity of the semantic representation of the same node in different views; while in the discriminator, MV-ACM discriminates whether the view is complementary by the topological structural similarity. Then we update the node’s semantic representation by aggregating neighborhoods information from the syncretic views. We conduct systematical experiments1 on six real-world networks from varied domains: AMiner, PPI, YouTube, Twitter, Amazon and Alibaba. Empirical results show that MV-ACM significantly outperforms the state-of-the-art approaches for both link prediction and node classification tasks.