Rationale discovery is defined as finding a subset of the input data that maximally supports the prediction of downstream tasks. In the context of graph machine learning, graph rationale is defined as identifying the critical subgraph in the given graph topology. In contrast to the rationale subgraph, the remaining subgraph is named the environment subgraph. Graph rationalization can enhance the model performance because the mapping between the graph rationale and the prediction label is viewed as invariant, by definition. To ensure the discriminative power of the extracted rationale subgraphs, a key technique named intervention is applied, whose core idea is that given changing environment subgraphs, the semantics from the rationale subgraph is invariant, which guarantees the correct prediction result. However, most, if not all, of the existing graph rationalization methods develop their intervention strategies on the graph level, which is coarse-grained. In this paper, we propose FIne-grained Graph rationalization (FIG). Our idea is driven by the self-attention mechanism, which provides rich interactions between input nodes. Based on that, FIG can achieve node-level and virtual node-level intervention. Our experiments involve 7 real-world datasets, and the proposed FIG shows significant performance advantages compared to 13 baseline methods.

Recent studies show that well-devised perturbations on graph structures or node features can mislead trained Graph Neural Network (GNN) models. However, these methods often overlook practical assumptions, over-rely on heuristics, or separate vital attack components. In response, we present GAIM, an integrated adversarial attack method conducted on a node feature basis while considering the strict black-box setting. Specifically, we define an adversarial influence function to assess the adversarial impact of node perturbations, thereby reframing the GNN attack problem into the adversarial influence maximization problem. In our approach, we unify the selection of the target node and the construction of feature perturbations into a single optimization problem, ensuring a unique and consistent feature perturbation for each target node. We use a surrogate model to transform this problem into a solvable linear programming task, streamlining the optimization process. Moreover, we extend our method to accommodate label- oriented attacks, broadening its applicability. Evaluations on five benchmark datasets across three popular models underscore the effectiveness of our method.

To deal with distribution shifts in graph data, various graph out-of-distribution (OOD) generalization techniques have been recently proposed. These methods often employ a two-step strategy that first creates augmented environments and subsequently identifies invariant subgraphs to improve generalizability. Nevertheless, this approach could be suboptimal from the perspective of consistency. First, the process of augmenting environments by altering the graphs while preserving labels may lead to graphs that are not realistic or meaningfully related to the origin distribution, thus lacking distribution consistency. Second, the extracted subgraphs are obtained from directly modifying graphs, and may not necessarily maintain a consistent predictive relationship with their labels, thereby impacting label consistency. In response to these challenges, we introduce an innovative approach that aims to enhance these two types of consistency for graph OOD generalization. We propose a modifier to obtain both augmented and invariant graphs in a unified manner. With the augmented graphs, we enrich the training data without compromising the integrity of label-graph relationships. The label consistency enhancement in our framework further preserves the supervision information in the invariant graph. We conduct extensive experiments on real-world datasets to demonstrate the superiority of our framework over other state-of-the-art baselines.

The Matrix Profile (MP), a versatile tool for time series data mining, has been shown effective in time series anomaly detection (TSAD). This paper delves into the problem of anomaly detection in multidimensional time series, a common occurrence in real-world applications. For instance, in a manufacturing factory, multiple sensors installed across the site collect time-varying data for analysis. The Matrix Profile, named for its role in profiling the matrix storing pairwise distance between subsequences of univariate time series, becomes complex in multidimensional scenarios. If the input univariate time series has$n$subsequences, the pairwise distance matrix is a$n\times n$matrix. In a multidimensional time series with$d$dimensions, the pairwise distance information must be stored in a$n\times n\times d$tensor. In this paper, we first analyze different strategies for condensing this tensor into a profile vector. We then investigate the potential of extending the MP to efficiently find k-nearest neighbors for anomaly detection. Finally, we benchmark the multidimensional MP against 19 baseline methods on 119 multidimensional TSAD datasets. The experiments covers three learning setups: unsupervised, supervised, and semi-supervised. MP is the only method that consistently delivers high performance across all setups.

Spatial-temporal forecasting systems play a crucial role in addressing numerous real-world challenges. In this paper, we investigate the potential of addressing spatial-temporal forecasting problems using general time series forecasting models, i.e., models that do not leverage the spatial relationships among the nodes. We propose a all-Multi-Layer Perceptron (all-MLP) time series forecasting architecture called RPMixer. The all-MLP architecture was chosen due to its recent success in time series forecasting benchmarks. Furthermore, our method capitalizes on the ensemble-like behavior of deep neural networks, where each individual block within the network behaves like a base learner in an ensemble model, particularly when identity mapping residual connections are incorporated. By integrating random projection layers into our model, we increase the diversity among the blocks' outputs, thereby improving the overall performance of the network. Extensive experiments conducted on the largest spatial-temporal forecasting benchmark datasets demonstrate that the proposed method outperforms 14 alternative methods.

Graph Transformers have garnered significant attention for learning graph-structured data, thanks to their superb ability to capture long-range dependencies among nodes. However, the quadratic space and time complexity hinders the scalability of Graph Transformers, particularly for large-scale recommendation. Here we propose an efficient Masked Graph Transformer, named MGFormer, capable of capturing all-pair interactions among nodes with a linear complexity. To achieve this, we treat all user/item nodes as independent tokens, enhance them with positional embeddings, and feed them into a kernelized attention module. Additionally, we incorporate learnable relative degree information to appropriately reweigh the attentions. Experimental results show the superior performance of our MGFormer, even with a single attention layer.

Contrastive Learning (CL) has shown promising performance in collaborative filtering. The key idea is to use contrastive loss to generate augmentation-invariant embeddings by maximizing the Mutual Information between different augmented views of the same instance. However, we empirically observe that existing CL models suffer from the dimensional collapse issue, where user/item embeddings only span a low-dimension subspace of the entire feature space. This suppresses other dimensional information and weakens the distinguishability of embeddings. Here we propose a non-contrastive learning objective, named nCL, which explicitly mitigates dimensional collapse of representations in collaborative filtering. Our nCL aims to achieve geometric properties of Alignment and Compactness on the embedding space. In particular, the alignment tries to push together representations of positive-related user-item pairs, while compactness tends to find the optimal coding length of user/item embeddings, subject to a given distortion. More importantly, our nCL does not require data augmentation nor negative sampling during training, making it scalable to large datasets compared to contrastive learning methods. Experimental results demonstrate the superiority of our nCL.

Network embedding plays an important role in a variety of social network applications. Existing network embedding methods, explicitly or implicitly, can be categorized into positional embedding (PE) methods or structural embedding (SE) methods. Specifically, PE methods encode the positional information and obtain similar embeddings for adjacent/close nodes, while SE methods aim to learn identical representations for nodes with the same local structural patterns, even if the two nodes are far away from each other. The disparate designs of the two types of methods lead to an apparent dilemma in that no embedding could perfectly capture both positional and structural information. In this paper, we seek to demystify the underlying relationship between positional embedding and structural embedding. We first point out that the positional embedding can produce the structural embedding with simple transformations, while the opposite direction cannot hold. Based on this finding, a novel network embedding model PACER is proposed, which optimizes the positional embedding with the help of random walk with restart (RWR) proximity distribution, and such positional embedding is then used to seamlessly obtain the structural embedding with simple transformations. Furthermore, two variants of PACER are proposed to handle node classification task on homophilic and heterophilic graphs. Extensive experiments on 17 datasets show that PACER achieves comparable or better performance than the state-of-the-arts.

Recommender systems (RSs) have gained widespread applications across various domains owing to the superior ability to capture users' interests. However, the complexity and nuanced nature of users' interests, which span a wide range of diversity, pose a significant challenge in delivering fair recommendations. In practice, user preferences vary significantly; some users show a clear preference toward certain item categories, while others have a broad interest in diverse ones. Even though it is expected that all users should receive high-quality recommendations, the effectiveness of RSs in catering to this disparate interest diversity remains under-explored.

Publishing and sharing data is crucial for the data mining community, allowing collaboration and driving open innovation. However, many researchers cannot release their data due to privacy regulations or fear of leaking confidential business information. To alleviate such issues, we propose the Time Series Synthesis Using the Matrix Profile (TSSUMP) method, where synthesized time series can be released in lieu of the original data. The TSSUMP method synthesizes time series by preserving similarity join information (i.e., Matrix Profile) while reducing the correlation between the synthesized and the original time series. As a result, neither the values for the individual time steps nor the local patterns (or shapes) from the original data can be recovered, yet the resulting data can be used for downstream tasks that data analysts are interested in. We concentrate on similarity joins because they are one of the most widely applied time series data mining routines across different data mining tasks. We test our method on a case study of ECG and gender masking prediction. In this case study, the gender information is not only removed from the synthesized time series, but the synthesized time series also preserves enough information from the original time series. As a result, unmodified data mining tools can obtain near-identical performance on the synthesized time series as on the original time series.

Time series data is ubiquitous across various domains such as finance, healthcare, and manufacturing, but their properties can vary significantly depending on the domain they originate from. The ability to perform Content-based Time Series Retrieval (CTSR) is crucial for identifying unknown time series examples. However, existing CTSR works typically focus on retrieving time series from a single domain database, which can be inadequate if the user does not know the source of the query time series. This limitation motivates us to investigate the CTSR problem in a scenario where the database contains time series from multiple domains. To facilitate this investigation, we introduce a CTSR benchmark dataset that comprises time series data from a variety of domains, such as motion, power demand, and traffic. This dataset is sourced from a publicly available time series classification dataset archive, making it easily accessible to researchers in the field. We compare several popular methods for modeling and retrieving time series data using this benchmark dataset. Additionally, we propose a novel distance learning model that outperforms the existing methods. Overall, our study highlights the importance of addressing the CTSR problem across multiple domains and provides a useful benchmark dataset for future research.

Time series classification is a widely studied problem in the field of time series data mining. Previous research has predominantly focused on scenarios where relevant or foreground subsequences have already been extracted, with each subsequence corresponding to a single label. However, real-world time series data often contain foreground subsequences that are intertwined with background subsequences. Successfully classifying these relevant subsequences requires not only distinguishing between different classes but also accurately identifying the foreground subsequences amidst the background. To address this challenge, we propose a novel subsequence classification method that represents each subsequence as an ego-network, providing crucial nearest neighbor information to the model. The ego-networks of all subsequences collectively form a time series subsequence graph, and we introduce an algorithm to efficiently construct this graph. Furthermore, we have demonstrated the significance of enforcing temporal consistency in the prediction of adjacent subsequences for the subsequence classification problem. To evaluate the effectiveness of our approach, we conducted experiments using 128 univariate and 30 multivariate time series datasets. The experimental results demonstrate the superior performance of our method compared to alternative approaches. Specifically, our method outperforms the baseline on 104 out of 158 datasets.

Time series discords are a useful primitive for time series anomaly detection, and the matrix profile is capable of capturing discord effectively. There exist many research efforts to improve the scalability of discord discovery with respect to the length of time series. However, there is surprisingly little work focused on reducing the time complexity of matrix profile computation associated with dimensionality of a multidimensional time series. In this work, we propose a sketch for discord mining among multi-dimensional time series. After an initial pre-processing of the sketch as fast as reading the data, the discord mining has runtime independent of the dimensionality of the original data. On several real world examples from water treatment and transportation, the proposed algorithm improves the throughput by at least an order of magnitude (50X) and only has minimal impact on the quality of the approximated solution. Additionally, the proposed method can handle the dynamic addition or deletion of dimensions inconsequential overhead. This allows a data analyst to consider “what-if” scenarios in real time while exploring the data.

Spatial-temporal graph neural networks (STGNNs) are promising in solving real-world spatial-temporal forecasting problems. Recognizing the inherent sequential relationship of spatial-temporal data, it is natural to explore the integration of boosting training mechanism to further enhance the performance of STGNNs. However, few studies have touched this research area. To bridge this gap, in this work, we propose spatial-temporal graph boosting networks, namely STGBN, which to the best of our knowledge is the first attempt to leverage gradient boosting for enhancing STGNNs. STGBN follows the general training procedure of conventional gradient boosting, but incorporates two distinctive designs to improve its efficiency in training on spatial-temporal graphs. Specifically, we design an incremental learning strategy that progressively includes spatial-temporal data into training. Additionally, we enforce an identical architecture for the base learner in all boosting iterations with each base learner inheriting from the one in the previous iteration. These designs facilitate rapid convergence of the base learner and expedite the overall training process. The base learner in STGBN is designed as a Transformer sandwich, which consists of two temporal Transformers on the top and bottom and a spatial Transformer in the middle. Structuring them in such a way helps the model capture long-range temporal dynamics, global spatial dependencies, and deep spatial-temporal interactions. We perform extensive spatial-temporal forecasting experiments on four spatial-temporal graph benchmarks. Promising results demonstrate the outstanding performance of STGBN against a wide range of state-of-the-art baseline models.

Imbalanced datasets are commonly observed in various real-world applications, presenting significant challenges in training classifiers. When working with large datasets, the imbalanced issue can be further exacerbated, making it exceptionally difficult to train classifiers effectively. To address the problem, over-sampling techniques have been developed to linearly interpolating data instances between minorities and their neighbors. However, in many real-world scenarios such as anomaly detection, minority instances are often dispersed diversely in the feature space rather than clustered together. Inspired by domain-agnostic data mix-up, we propose generating synthetic samples iteratively by mixing data samples from both minority and majority classes. It is non-trivial to develop such a framework, the challenges include source sample selection, mix-up strategy selection, and the coordination between the underlying model and mix-up strategies. To tackle these challenges, we formulate the problem of iterative data mix-up as a Markov decision process (MDP) that maps data attributes onto an augmentation strategy. To solve the MDP, we employ an actor-critic framework to adapt the discrete-continuous decision space. This framework is utilized to train a data augmentation policy and design a reward signal that explores classifier uncertainty and encourages performance improvement, irrespective of the classifier's convergence. We demonstrate the effectiveness of our proposed framework through extensive experiments conducted on seven publicly available benchmark datasets using three different types of classifiers. The results of these experiments showcase the potential and promise of our framework in addressing imbalanced datasets with diverse minorities.

A Content-based Time Series Retrieval (CTSR) system is an information retrieval system for users to interact with time series emerged from multiple domains, such as finance, healthcare, and manufacturing. For example, users seeking to learn more about the source of a time series can submit the time series as a query to the CTSR system and retrieve a list of relevant time series with associated metadata. By analyzing the retrieved metadata, users can gather more information about the source of the time series. Because the CTSR system is required to work with time series data from diverse domains, it needs a high-capacity model to effectively measure the similarity between different time series. On top of that, the model within the CTSR system has to compute the similarity scores in an efficient manner as the users interact with the system in real-time. In this paper, we propose an effective and efficient CTSR model that outperforms alternative models, while still providing reasonable inference runtimes. To demonstrate the capability of the proposed method in solving business problems, we compare it against alternative models using our in-house transaction data. Our findings reveal that the proposed model is the most suitable solution compared to others for our transaction data problem.

A foundation model is a machine learning model trained on a large and diverse set of data, typically using self-supervised learning-based pre-training techniques, that can be adapted to various downstream tasks. However, current research on time series pre-training has predominantly focused on models trained exclusively on data from a single domain. As a result, these models possess domain-specific knowledge that may not be easily transferable to time series from other domains. In this paper, we aim to develop an effective time series foundation model by leveraging unlabeled samples from multiple domains. To achieve this, we repurposed the publicly available UCR Archive and evaluated four existing self-supervised learning-based pre-training methods, along with a novel method, on the datasets. We tested these methods using four popular neural network architectures for time series to understand how the pre-training methods interact with different network designs. Our experimental results show that pre-training improves downstream classification tasks by enhancing the convergence of the fine-tuning process. Furthermore, we found that the proposed pre-training method, when combined with the Transformer, outperforms the alternatives. The proposed method outperforms or achieves equal performance compared to the second best method in ~93% of downstream tasks.

Federated Learning (FL) enables multiple clients to collaboratively learn a machine learning model without exchanging their own local data. In this way, the server can exploit the computational power of all clients and train the model on a larger set of data samples among all clients. Although such a mechanism is proven to be effective in various fields, existing works generally assume that each client preserves sufficient data for training. In practice, however, certain clients can only contain a limited number of samples (i.e., few-shot samples). For example, the available photo data taken by a specific user with a new mobile device is relatively rare. In this scenario, existing FL efforts typically encounter a significant performance drop on these clients. Therefore, it is urgent to develop a few-shot model that can generalize to clients with limited data under the FL scenario. In this paper, we refer to this novel problem as federated few-shot learning. Nevertheless, the problem remains challenging due to two major reasons: the global data variance among clients (i.e., the difference in data distributions among clients) and the local data insufficiency in each client (i.e., the lack of adequate local data for training). To overcome these two challenges, we propose a novel federated few-shot learning framework with two separately updated models and dedicated training strategies to reduce the adverse impact of global data variance and local data insufficiency. Extensive experiments on four prevalent datasets that cover news articles and images validate the effectiveness of our framework compared with the state-of-the-art baselines.

The huge volume of emerging graph datasets has become a double-bladed sword for graph machine learning. On the one hand, it empowers the success of a myriad of graph neural networks (GNNs) with strong empirical performance. On the other hand, training modern graph neural networks on huge graph data is computationally expensive. How to distill the given graph dataset while retaining most of the trained models' performance is a challenging problem. Existing efforts try to approach this problem by solving meta-learning-based bilevel optimization objectives. A major hurdle lies in that the exact solutions of these methods are computationally intensive and thus, most, if not all, of them are solved by approximate strategies which in turn hurt the distillation performance. In this paper, inspired by the recent advances in neural network kernel methods, we adopt a kernel ridge regression-based meta-learning objective which has a feasible exact solution. However, the computation of graph neural tangent kernel is very expensive, especially in the context of dataset distillation. As a response, we design a graph kernel, named LiteGNTK, tailored for the dataset distillation problem which is closely related to the classic random walk graph kernel. An effective model named Kernel rıdge regression-based graph Dataset Distillation (KIDD) and its variants are proposed. KIDD shows nice efficiency in both the forward and backward propagation processes. At the same time, KIDD shows strong empirical performance over 7 real-world datasets compared with the state-of-the-art distillation methods. Thanks to the ability to find the exact solution of the distillation objective, the learned training graphs by KIDD can sometimes even outperform the original whole training set with as few as 1.65% training graphs.

Collaborative Filtering (CF) has been successfully used to help users discover the items of interest. Nevertheless, existing CF methods suffer from noisy data issue, which negatively impacts the quality of recommendation. To tackle this problem, many prior studies leverage adversarial learning to regularize the representations of users/items, which improves both generalizability and robustness. Those methods often learn adversarial perturbations and model parameters under min-max optimization framework. However, there still have two major drawbacks: 1) Existing methods lack theoretical guarantees of why adding perturbations improve the model generalizability and robustness; 2) Solving min-max optimization is time-consuming. In addition to updating the model parameters, each iteration requires additional computations to update the perturbations, making them not scalable for industry-scale datasets.

Graph Neural Networks (GNNs) have achieved state-of-the-art performance in recommender systems. Nevertheless, the process of searching and ranking from a large item corpus usually requires high latency, which limits the widespread deployment of GNNs in industry-scale applications. To address this issue, many methods compress user/item representations into the binary embedding space to reduce space requirements and accelerate inference. Also, they use the Straight-through Estimator (STE) to prevent vanishing gradients during back-propagation. However, the STE often causes the gradient mismatch problem, leading to sub-optimal results.

Transformer and its variants are a powerful class of architectures for sequential recommendation, owing to their ability of capturing a user’s dynamic interests from their past interactions. Despite their success, Transformer-based models often require the optimization of a large number of parameters, making them difficult to train from sparse data in sequential recommendation. To address the problem of data sparsity, previous studies have utilized self-supervised learning to enhance Transformers, such as pre-training embeddings from item attributes or contrastive data augmentations. However, these approaches encounter several training issues, including initialization sensitivity, manual data augmentations, and large batch-size memory bottlenecks.

Time series anomaly detection is a challenging task with a wide range of real-world applications. Due to label sparsity, training a deep anomaly detector often relies on unsupervised approaches. Recent efforts have been devoted to time series domain adaptation to leverage knowledge from similar domains. However, existing solutions may suffer from negative knowledge transfer on anomalies due to their diversity and sparsity. Motivated by the empirical study of context alignment between two domains, we aim to transfer knowledge between two domains via adaptively sampling context information for two domains. This is challenging because it requires simultaneously modeling the complex in-domain temporal dependencies and cross-domain correlations while exploiting label information from the source domain. To this end, we propose a framework that combines context sampling and anomaly detection into a joint learning procedure. We formulate context sampling into the Markov decision process and exploit deep reinforcement learning to optimize the time series domain adaptation process via context sampling and design a tailored reward function to generate domain-invariant features that better align two domains for anomaly detection. Experiments on three public datasets show promise for knowledge transfer between two similar domains and two entirely different domains.

Graph Neural Networks (GNNs) have achieved impressive performance in collaborative filtering. However, recent studies show that GNNs tend to yield inferior performance when the distributions of training and test data are not aligned well. Moreover, training GNNs often requires optimizing non-convex neural networks with an abundance of local and global minima, which may differ widely in their performance at test time. Thus, it is essential to develop an optimization strategy that can choose the minima carefully, which can yield strong generalization performance on unseen data. Here we propose an effective training schema, called gSAM, under the principle that theflatter minima has a better generalization ability than thesharper ones. To achieve this goal, gSAM regularizes the flatness of the weight loss landscape by forming a bi-level optimization: the outer problem conducts the standard model training while the inner problem helps the model jump out of the sharp minima. Experimental results show the superiority of our gSAM.

Graph Neural Networks (GNNs) have shown great power in learning node representations on graphs. However, they may inherit historical prejudices from training data, leading to discriminatory bias in predictions. Although some work has developed fair GNNs, most of them directly borrow fair representation learning techniques from non-graph domains without considering the potential problem of sensitive attribute leakage caused by feature propagation in GNNs. However, we empirically observe that feature propagation could vary the correlation of previously innocuous non-sensitive features to the sensitive ones. This can be viewed as a leakage of sensitive information which could further exacerbate discrimination in predictions. Thus, we design two feature masking strategies according to feature correlations to highlight the importance of considering feature propagation and correlation variation in alleviating discrimination. Motivated by our analysis, we propose Fair View Graph Neural Network (FairVGNN) to generate fair views of features by automatically identifying and masking sensitive-correlated features considering correlation variation after feature propagation. Given the learned fair views, we adaptively clamp weights of the encoder to avoid using sensitive-related features. Experiments on real-world datasets demonstrate that FairVGNN enjoys a better trade-off between model utility and fairness.

Graph neural networks (GNNs) are deep learning models designed specifically for graph data, and they typically rely on node features as the input to the first layer. When applying such a type of network on the graph without node features, one can extract simple graph-based node features (e.g., number of degrees) or learn the input node representations (i.e., embeddings) when training the network. While the latter approach, which trains node embeddings, more likely leads to better performance, the number of parameters associated with the embeddings grows linearly with the number of nodes. It is therefore impractical to train the input node embeddings together with GNNs within graphics processing unit (GPU) memory in an end-to-end fashion when dealing with industrial-scale graph data. Inspired by the embedding compression methods developed for natural language processing (NLP) tasks, we develop a node embedding compression method where each node is compactly represented with a bit vector instead of a floating-point vector. The parameters utilized in the compression method can be trained together with GNNs. We show that the proposed node embedding compression method achieves superior performance compared to the alternatives.

Transformer-based sequential recommenders are very powerful for capturing both short-term and long-term sequential item dependencies. This is mainly attributed to their unique self-attention networks to exploit pairwise item-item interactions within the sequence. However, real-world item sequences are often noisy, which is particularly true for implicit feedback. For example, a large portion of clicks do not align well with user preferences, and many products end up with negative reviews or being returned. As such, the current user action only depends on a subset of items, not on the entire sequences. Many existing Transformer-based models use full attention distributions, which inevitably assign certain credits to irrelevant items. This may lead to sub-optimal performance if Transformers are not regularized properly.

There has been an explosion of interest in designing various Knowledge Graph Neural Networks (KGNNs), which achieve state-of-the-art performance and provide great explainability for recommendation. The promising performance is mainly resulting from their capability of capturing high-order proximity messages over the knowledge graphs. However, training KGNNs at scale is challenging due to the high memory usage. In the forward pass, the automatic differentiation engines (e.g., TensorFlow/PyTorch) generally need to cache all intermediate activation maps in order to compute gradients in the backward pass, which leads to a large GPU memory footprint. Existing work solves this problem by utilizing multi-GPU distributed frameworks. Nonetheless, this poses a practical challenge when seeking to deploy KGNNs in memory-constrained environments, especially for industry-scale graphs.

The matrix profile is an effective data mining tool that provides similarity join functionality for time series data. Since the introduction of the matrix profile five years ago, multiple efforts have been made to speed up the computation with approximate joins; however, the majority of these efforts only focus on self-joins. In this work, we show that it is possible to efficiently perform approximate inter-time series similarity joins with error bounded guarantees by creating a compact “dictionary” representation of time series. Using the dictionary representation instead of the original time series, we are able to improve the throughput of an anomaly mining system by at least 20X, with essentially no decrease in accuracy. As a side effect, the dictionaries also summarize the time series in a semantically meaningful way and can provide intuitive and actionable insights.

Graph Neural Networks (GNNs) provide a class of powerful architectures that are effective for graph-based collaborative filtering. Nevertheless, GNNs are known to be vulnerable to adversarial perturbations. Adversarial training is a simple yet effective way to improve the robustness of neural models. For example, many prior studies inject adversarial perturbations into either node features or hidden layers of GNNs. However, perturbing graph structures has been far less studied in recommendations.

Graph Neural Networks (GNNs) have emerged as powerful tools for collaborative filtering. A key challenge of recommendations is to distill long-range collaborative signals from user-item graphs. Typically, GNNs generate embeddings of users/items by propagating and aggregating the messages between local neighbors. Thus, the ability of GNNs to capture long-range dependencies heavily depends on their depths. However, simply training deep GNNs has several bottleneck effects, e.g., over-fitting & over-smoothing, which may lead to unexpected results if GNNs are not well regularized.

Today’s cyber-world is vastly multivariate. Metrics collected at extreme varieties demand multivariate algorithms to properly detect anomalies. However, forecast-based algorithms, as widely proven approaches, often perform sub-optimally or inconsistently across datasets. A key common issue is they strive to be one-size-fits-all but anomalies are distinctive in nature. We propose a method that tailors to such distinction. Presenting FMUAD - a Forecast-based, Multi-aspect, Unsupervised Anomaly Detection framework. FMUAD explicitly and separately captures the signature traits of anomaly types - spatial change, temporal change and correlation change - with independent modules. The modules then jointly learn an optimal feature representation, which is highly flexible and intuitive, unlike most other models in the category. Extensive experiments show our FMUAD framework consistently outperforms other state-of-the-art forecast-based anomaly detectors.

Fashion is more than Paris runways. Fashion is about how people express their interests, identity, mood, and cultural influences. Given an inventory of candidate garments from different categories, how to assemble them together would most improve their fashionability? This question presents an intriguing visual recommendation challenge to automatically create capsule wardrobes. Capsule wardrobe generation is a complex combinatorial problem that requires the understanding of how multiple visual items interact. The generative process often needs fashion experts to manually tease the combinations out, making it hard to scale.

Graph Convolutional Networks (GCNs) are powerful for collaborative filtering. The key component of GCNs is to explore neighborhood aggregation mechanisms to extract high-level representations of users and items. However, real-world user-item graphs are often incomplete and noisy. Aggregating misleading neighborhood information may lead to sub-optimal performance if GCNs are not regularized properly. Also, the real-world user-item graphs are often sparse and low rank. These two intrinsic graph properties are widely used in shallow matrix completion models, but far less studied in graph neural models. Here we propose Structured Graph Convolutional Networks (SGCNs) to enhance the performance of GCNs by exploiting graph structural properties of sparsity and low rank. To achieve sparsity, we attach each layer of a GCN with a trainable stochastic binary mask to prune noisy and insignificant edges, resulting in a clean and sparsified graph. To preserve its low-rank property, the nuclear norm regularization is applied. We jointly learn the parameters of stochastic binary masks and original GCNs by solving a stochastic binary optimization problem. An unbiased gradient estimator is further proposed to better backpropagate the gradients of binary variables. Experimental results demonstrate that SGCNs achieve better performance compared with the state-of-the-art GCNs.

Clustering is one of the most prominent topics in machine learning. A multitude of clustering methods have been proposed, among which the spectral clustering has attracted much attention. However, in practice, spectral clustering is highly sensitive to noise data and a post-processing step (e.g., k-means for eigenvectors) is often required to obtain clustering indicators, which may be not optimal. Also, it does not scale well to large-scale data due to its eigen-decomposition procedures.Here we propose a structure-aware clustering model to address those issues. To achieve our goal, a high-quality affinity matrix is extracted from the original noisy data by a sparse additive decomposition, which is used to approximate the ideal clustering structure. We then jointly learn the high-quality affinity matrix as well as the spectral embedding in a unified model- thus, being robust to noise and obtaining the optimal clustering indicators without any post-processing steps. We further improve the clustering stability by considering the Laplacian eigengap of the affinity matrix. We show that the larger the Laplacian eigengap, the more stable the clustering results. We introduce a speedup strategy to effectively compute eigenvectors of large matrices. Experimental results demonstrate that the proposed model outperforms existing approaches for noisy data.

Tag and item recommendation are both important tasks in recommender systems. Item recommendation, learning user-item rating patterns, helps users to discover a list of items that are more likely to match their interests. On the other hand, tag recommendation, exploiting relationships of user-tag-item, assists users to find a personalized list of tags that can precisely describe their experience on items. Existing methods often consider these two tasks independently. However, these two tasks are inherently related because they both use information from same groups of users and items. In other words, it is the same users who give ratings and annotate tags to the same set of items. Tags may explain why a user purchase an item. For instance, customers on Yelp not only give a five star rating for a restaurant but also leave comments/tags to express their positive sentiments. Users with similar interests might give similar ratings and post similar tags to certain items. The co-occurrence of users, items and tags motivates us to jointly perform item and tag recommendation to better model users' behavior by leveraging their personalized items and tags. In this work, we propose a cross-domain learning method that seamlessly integrates tag and item recommendation into one unified framework. In particular, we utilize the knowledge transferred between two tasks to collaboratively rank items and tags for users through coupled tensor-matrix factorization. We also provide a principled way to incorporate additional auxiliary information about users (e.g., users' relationships in a social network) and items (e.g., descriptions of items in Wikipedia) to improve recommendation accuracy for both tasks. Empirical results demonstrate the effectiveness of our proposed model.

Tensor completion has been successfully applied to many real-world applications. In a wide variety of situations, data utilized in many learning tasks are of high dimensions, usually extracted from multiple heterogeneous sources. Therefore, data can be represented by a primary tensor and multiple matrices generated from multi-view side information or metadata. Joint analysis of tensors and matrices has great potential to gain better understanding of the underlying relationships among these multiple heterogeneous sources. The existing tensor completion methods, which recover the missing elements of a partially known tensor with single view side information, can yield interpretable results for large-scale datasets. However, their limitations up to now are lack of modeling multi-view heterogeneous data and suitably learning the low-rank property of tensor. In this study, we fill this gap by developing a novel collective tensor completion method, which tightly fuses multi-view heterogeneous data sources. Our method exploits special common latent structures from the primary tensor and multiple side matrices through coupled tensor-matrix decomposition, in which the common latent structures can compactly represent all the data. In addition, rank estimation of a tensor is a challenging task due to its discrete nature. Instead of approximating the rank by widely used trace norm or nuclear norm, we directly utilize Schatten p-norm on the latent structures to better approximate the rank and to enhance its robustness to noise.

Cross-domain recommendation has attracted growing interests given their simplicity and effectiveness. In the cross-domain scenarios, we may improve predictive accuracy in one domain by transferring knowledge from the other, which alleviates the data sparsity issue. However, the relatedness of these domains can be exploited by a malicious party to launch data poisoning attacks. Here we study the vulnerability of cross-domain recommendation under data poisoning attacks. We show that data poisoning attacks can be formulated as a bilevel optimization problem. Our experimental results show that cross-domain system can be compromised under attacks, highlighting the need for countermeasures against data poisoning attacks in cross-domain recommendation.

Contextual factors such as time, location, or tag, can affect user preferences for a particular item. Context-aware recommendations are thus critical to improve both quality and explainability of recommender systems, compared to traditional recommendations that are solely based on user-item interactions. Tensor factorization machines have achieved the state-of-the-art performance due to their capability of integrating users, items, and contextual factors in one unify way. However, few work has focused on the robustness of a context-aware recommender system. Improving the robustness of a tensor-based model is challenging due to the sparsity of the observed tensor and the multi-linear nature of tensor factorization. In this paper, we propose ATF, a model that combines tensor factorization and adversarial learning for context-aware recommendations. Doing so allows us to reap the benefits of tensor factorization, while enhancing the robustness of a recommender model, and thus improves its eventual performance. Empirical studies on two real-world datasets show that the proposed method outperforms standard tensor-based methods.

Modeling the behaviors of drug-target-disease interactions is crucial in the early stage of drug discovery and holds great promise for precision medicine and personalized treatments. The growing availability of new types of data on the internet brings great opportunity of learning a more comprehensive relationship among drugs, targets, and diseases. However, existing methods often consider drug-target interactions or drug-disease interactions separately, which ignores the dependencies among these three entities. Also, many of them cannot directly incorporate rich heterogeneous information from diverse sources.

Link prediction in dynamic networks is an important task with many real-life applications in different domains, such as social networks, cyber-physical systems, and bioinformatics. There are two key processes in dynamic networks: network structural evolution and network temporal evolution, where the former represents interdependency between entities and their neighbors in the network at each timestamp, while the latter captures the evolving behavior of the entire network from the current timestamp to the next. Structural evolution generally assumes that a node is more likely to co-evolve with its neighbors in the near future. Temporal evolution focuses on the trend of network evolution as a whole, based on the accumulation of historical data. It is thus essential to use characteristics of both structural and temporal evolutions to emulate complex behaviors of dynamic networks. However, very few existing work considered both processes. In addition, real-life networks are often very sparse with limited observed links. A missing link between two nodes does not always imply that the two nodes do not have a relation in reality, especially when they share many common neighbors. Most existing methods only focus on the first-order proximity of networks, which is usually insufficient to capture the relationships among nodes.

Personalized treatments and targeted therapies are the most promising approaches to treat complex diseases, especially for cancer. However, drug resistance is often acquired after treatments. To overcome or reduce drug resistance, treatments using drug combinations have been actively investigated in the literature. Existing methods mainly focus on chemical properties of drugs for potential combination therapies without considering relationships among different diseases. Also, they often do not consider the rich knowledge of drugs and diseases, which can enhance the prediction of drug combinations. This motivates us to develop a new computational method that can predict the beneficial drug combinations. We propose DrugCom, a tensor-based framework for computing drug combinations across different diseases by integrating multiple heterogeneous data sources of drugs and diseases. DrugCom first constructs a primary third-order tensor (i.e., drug×drug ×disease) and several similarity matrices from multiple data sources regarding drugs (e.g., chemical structure) and diseases (e.g., disease phenotype). DrugCom then formulates an objective function, which simultaneously factorizes coupled tensor and matrices to reveal the molecular mechanisms of drug synergy. We adopt the alternating direction method of multipliers algorithm to effectively solve the optimization problem. Extensive experimental studies using real-world datasets demonstrate superior performance of DrugCom.

In addition to the sparse user-item (U-I) matrix, an increasing number of current recommender systems seek to improve performance by exploiting extra heterogeneous data sources (e.g., online social networks). Such rich side sources can provide very useful information about users' personal behaviors and items' properties, therefore can significantly benefit recommender systems. Most existing work can only incorporate a single side source of users or items. In many real-life applications, however, there exist multiple side sources for both users and items, which may be constructed from multiple domains. To incorporate multiple sources, we propose a flexible and robust algorithm called MSUI (Learning Multiple Similarities of Users and Items). The key idea of MSUI is to simultaneously capture the associations between latent factors in the U-I matrix and multiple side sources of users and items through joint nonnegative matrix factorization. MSUI has several advantages over existing methods. First, it seamlessly integrates information from the U-I matrix and information from multiple side sources. Second, by incorporating multiple sources that are usually in complementary, it's more robust to noise contained in each source. Finally, it provides a unified framework that can be easily extended to incorporate additional data sources.