Distribution shifts on graphs—the discrepancies in data distribution between training and employing a graph machine learning model—are ubiquitous and often unavoidable in real-world applications. These shifts may severely deteriorate model performance, posing significant challenges for reliable graph machine learning. In recent years, there has been a surge in research on graph machine learning specifically designed to tackle such distribution shifts, aiming to train models to achieve satisfactory performance on Out-of-Distribution (OOD) test data. This survey provides an up-to-date and forward-looking review of deep graph learning under distribution shifts. We categorize the field into three primary scenarios: graph OOD generalization, training-time graph OOD adaptation, and test-time graph OOD adaptation. We begin by formally formulating the problems and discussing various types of distribution shifts that can affect graph learning, such as covariate shifts and concept shifts. To provide a structured understanding of the literature, we introduce a systematic taxonomy that classifies existing methods into model-centric and data-centric approaches, investigating the techniques used in each category. We also summarize commonly used datasets in this research area to facilitate further investigation. Finally, we point out promising research directions and the corresponding challenges to encourage further study in this vital domain. Additionally, we provide a continuously updated reading list at https://github.com/kaize0409/Awesome-Graph-OOD .

Large pre-trained models have achieved notable success across a range of downstream tasks. However, recent research shows that a type of adversarial attack (i.e., backdoor attack) can manipulate the behavior of machine learning models through contaminating their training dataset, posing significant threat in the real-world application of large pre-trained model, especially for those customized models. Therefore, addressing the unique challenges for exploring vulnerability of pre-trained models is of paramount importance. Through empirical studies on the capability for performing backdoor attack in large pre-trained models (e.g., ViT), we find the following unique challenges of attacking large pre-trained models: 1) the inability to manipulate or even access large training datasets, and 2) the substantial computational resources required for training or fine-tuning these models. To address these challenges, we establish new standards for an effective and feasible backdoor attack in the context of large pre-trained models. In line with these standards, we introduce our EDT model, an Efficient, Data-free, Training-free backdoor attack method. Inspired by model editing techniques, EDT injects an editing-based lightweight codebook into the backdoor of large pre-trained models, which replaces the embedding of the poisoned image with the target image without poisoning the training dataset or training the victim model. Our experiments, conducted across various pre-trained models such as ViT, CLIP, BLIP, and stable diffusion, and on downstream tasks including image classification, image captioning, and image generation, demonstrate the effectiveness of our method. Our code is available at https://github.com/donglgcn/Editing/

Causality lays the foundation for the trajectory of our world. Causal inference (CI), which aims to infer intrinsic causal relations among variables of interest, has emerged as a crucial research topic. Nevertheless, the lack of observation of important variables (e.g., confounders, mediators, exogenous variables, etc.) severely compromises the reliability of CI methods. The issue may arise from the inherent difficulty in measuring the variables. Additionally, in observational studies where variables are passively recorded, certain covariates might be inadvertently omitted by the experimenter. Depending on the type of unobserved variables and the specific CI task, various consequences can be incurred if these latent variables are carelessly handled, such as biased estimation of causal effects, incomplete understanding of causal mechanisms, lack of individual-level causal consideration, etc. In this survey, we provide a comprehensive review of recent developments in CI with latent variables. We start by discussing traditional CI techniques when variables of interest are assumed to be fully observed. Afterward, under the taxonomy of circumvention and inference-based methods, we provide an in-depth discussion of various CI strategies to handle latent variables, covering the tasks of causal effect estimation, mediation analysis, counterfactual reasoning, and causal discovery. Furthermore, we generalize the discussion to graph data where interference among units may exist. Finally, we offer fresh aspects for further advancement of CI with latent variables, especially new opportunities in the era of large language models (LLMs).

Video anomaly detection (VAD) aims at identifying unusual behaviors from videos. Most of the existing video anomaly detection methods can achieve promising performance in the scenarios where training and test samples are drawn from the same distribution. In real-world situation, however, it is intractable to collect and label sufficient training video samples that cover many possible test scenarios, and existing methods demonstrate limited generalization ability. Focusing on this issue, we present the few-shot cross-domain video anomaly detection (FC-VAD) problem, which aims to adapt anomaly detection model to target samples, with access to only a few target video frames. To solve the FC-VAD problem, we propose an adaptive video anomaly detection framework named Ada-VAD, which contains a pretraining stage and an adaptation stage. In the pretraining stage, we synthesize abnormal samples and design a self-supervision based prediction task to pretrain a domain invariant model. In the adaptation stage, we adapt the pre-trained model to target domain with few-shot samples by mitigating the distribution shift with an adversarial training approach. We conduct extensive experiments on three benchmark datasets, and results show that our Ada-VAD approach outperforms the state-of-the-art VAD methods in most cases. Our code is available at https://github.com/donglgcn/ADA-VAD

Geolocating precise locations from images presents a challenging problem in computer vision and information retrieval. Traditional methods typically employ either classification-dividing the Earth's surface into grid cells and classifying images accordingly, or retrieval-identifying locations by matching images with a database of image-location pairs. However, classification-based approaches are limited by the cell size and cannot yield precise predictions, while retrieval-based systems usually suffer from poor search quality and inadequate coverage of the global landscape at varied scale and aggregation levels. To overcome these drawbacks, we present Img2Loc, a novel system that redefines image geolocalization as a text generation task. This is achieved using cutting-edge large multi-modality models (LMMs) like GPT-4V or LLaVA with retrieval augmented generation. Img2Loc first employs CLIP-based representations to generate an image-based coordinate query database. It then uniquely combines query results with images itself, forming elaborate prompts customized for LMMs. When tested on benchmark datasets such as Im2GPS3k and YFCC4k, Img2Loc not only surpasses the performance of previous state-of-the-art models but does so without any model training. A video demonstration of the system can be accessed via this link https://drive.google.com/file/d/16A6A-mc7AyUoKHRH3_WBRToRC13sn7tU/view?usp=sharing

As AI systems have obtained significant performance to be deployed widely in our daily lives and human society, people both enjoy the benefits brought by these technologies and suffer many social issues induced by these systems. To make AI systems good enough and trustworthy, plenty of researches have been done to build guidelines for trustworthy AI systems. Machine learning is one of the most important parts of AI systems, and representation learning is the fundamental technology in machine learning. How to make representation learning trustworthy in real-world application, e.g., cross domain scenarios, is very valuable and necessary for both machine learning and AI system fields. Inspired by the concepts in trustworthy AI, we proposed the first trustworthy representation learning across domains framework, which includes four concepts, i.e., robustness, privacy, fairness, and explainability, to give a comprehensive literature review on this research direction. Specifically, we first introduce the details of the proposed trustworthy framework for representation learning across domains. Second, we provide basic notions and comprehensively summarize existing methods for the trustworthy framework from four concepts. Finally, we conclude this survey with insights and discussions on future research directions.

The era of big data has witnessed an increasing availability of observational data from mobile and social networking, online advertising, web mining, healthcare, education, public policy, marketing campaigns, and so on, which facilitates the development of causal effect estimation. Although significant advances have been made to overcome the challenges in the academic area, such as missing counterfactual outcomes and selection bias, they only focus on source-specific and stationary observational data, which is unrealistic in most industrial applications. In this paper, we investigate a new industrial problem of causal effect estimation from incrementally available observational data and present three new evaluation criteria accordingly, including extensibility, adaptability, and accessibility. We propose a Continual Causal Effect Representation Learning method for estimating causal effects with observational data, which are incrementally available from non-stationary data distributions. Instead of having access to all seen observational data, our method only stores a limited subset of feature representations learned from previous data. Combining selective and balanced representation learning, feature representation distillation, and feature transformation, our method achieves the continual causal effect estimation for new data without compromising the estimation capability for original data. Extensive experiments demonstrate the significance of continual causal effect estimation and the effectiveness of our method.

Representation learning (RL) aims to extract latent features from various types of data and then facilitate a wide range of downstream data analytics tasks, such as classification, clustering, outlier detection, recommender systems, etc. Prior efforts on RL in the past decades mainly focus on developing models to largely retain useful information (e.g., discriminative patterns, semantic knowledge) from data while discard redundant and noisy information. In the data mining and machine learning communities, some recent efforts attempt to encourage both task-oriented performance and trustworthiness of the model, suggesting a pathway towards trustworthy representation learning (TRL). Although trustworthiness has been increasingly discussed from different perspectives (e.g., fairness, explainability, robustness), the intertwined connections between representation learning and trustworthiness have not been formally discussed and clearly revealed yet. In this Blue Sky vision paper, for the first time, we present a conceptual framework that illustrates how to characterize trustworthiness in representation learning, discuss the research challenges, and point out future research opportunities.

The proliferation of observational data demands the development of statistical methods for causal inference. Many widely used causal inference methods are based on the propensity score. When estimating the propensity score, one essential question is which covariates should be included in the model. In this paper, we propose a deep adaptive variable selection based propensity score method (DAVSPS) by using representation learning and adaptive group LASSO. The key idea of DAFSPS is to combine the data-driven learning capability of representation learning and variable selection consistency of adaptive group LASSO to improve the estimation of the propensity score by selecting confounders and adjustment variables while removing instrumental and spurious variables. We also provide a detailed theoretical analysis to prove the variable selection consistency of DAVSPS. We evaluate the performance of our method on simulated data to demonstrate its superiority over state-of-the-art methods and apply it to real data.

Short text clustering is a challenging unsupervised learning task which requires a complex representation of each document to effectively model the semantics and syntactic structure of the text. Existing works have attempted to tackle this challenging task by incorporating additional information to the model, such as number of clusters, number of datapoints in each clusters, the distribution of the input data, and more. Unlike previous approaches, we propose to exploit an auxiliary dataset that is fully labeled to augment the quality of the learned representations. We also define the problem as cross domain clustering (XDC), which leverages adversarial learning to train an adaptive clustering model across text domains. Specifically, XDC jointly exploits a labeled source domain and an unlabeled target domain during model training. Owing to domain adversarial learning, the distribution shift across source and target domains could be mitigated. Moreover, XDC is implemented as a linkage-based clustering approach using graphs, which is agnostic of the number of clusters. We evaluate our XDC framework on three text datasets, and results show that it outperforms the state-of-the-art text clustering methods in most cases. Ablation studies and qualitative analysis also demonstrate the effectiveness of our framework.

With the increasing growth of data and the ability of learning with them, machine learning models are adopted in various domains. However, few of machine learning models are able to reason their prediction, which limits their further applications in real-world tasks. With the potential to address this dilemma, model interpretation has become an important research topic because of the ability to provide the underlying reasons for model predictions at the feature level or concept level. Model interpretation at the concept level focuses on exploring the roles of concepts in model prediction, which enables more compact and understandable interpretations. Concept-level model interpretation requires the identification of the concepts that contribute to model prediction and the exploration of the rules underneath these concepts. To achieve the two objectives, we propose a Concept-level Model Interpretation framework (CMIC) from the perspective of causality. CMIC can automatically detect concepts in data and discover the causal relation between the detected concepts and the model's predicted labels. Furthermore, CMIC ranks the contributions of concepts by their causal effect on the model prediction, reflecting the detected concepts’ importance. We evaluate the proposed CMIC framework on both synthetic and real-world datasets to demonstrate the quality of the provided interpretation.

Inspired by the success of transformers in natural language processing, vision transformers have been proposed to address a wide range of computer vision tasks, such as image classification, object detection and image segmentation, and they have achieved very promising performance. However, the robustness of vision transformers has been relatively under-explored. Recent studies have revealed that pre-trained vision transformers are also vulnerable to white-box adversarial attacks on the downstream image classification task. The adversarial attacks (e.g., FGSM and PGD) designed for convolutional neural networks (CNNs) can also cause severe performance drop for vision transformers. In this paper, we evaluate the robustness of vision transformers fine-tuned with the off-the-shelf methods under adversarial attacks on CIFAR-10 and CIFAR-100. We further propose a data-augmented virtual adversarial training approach called MixVAT, which is able to enhance the robustness of pre-trained vision transformers against adversarial attacks on the downstream tasks with the unlabelled data. Extensive results on multiple datasets demonstrate the superiority of our approach over baselines on adversarial robustness, without compromising generalization ability of the model.

Brook trout (Salvelinus fontinalis) is a freshwater fish species of ecological, economic, and cultural importance in eastern North America. Estimating the abundance, movement, and survival of brook trout in the wild is an important task for environmental management, and current methods often involve physical tagging or collection of DNA samples for each of the fish as their unique identifier. However, this process is expensive and inefficient. Meanwhile, although deep learning methods have proven effective for individual recognition of humans, it remains challenging to apply this system to wildlife biology due to fewer available images, different biometric patterns, and relatively poor image quality. In this paper, we develop a framework to automate the process of individual recognition of brook trout. Distinguished from simply adopting traditional feature descriptors (e.g., SIFT and HOG) or using deep neural networks on the raw images, our framework utilizes multiple modalities consisting of the region of interest and gray-scaled pigmentation patterns. We use these multiple modality features in a Convolutional Neural network to generate feature vectors as fish descriptors. These descriptors are then used to distinguish individual brook trout by ranking their relative distance in latent space. Our experimental framework demonstrates better results than baseline methods such as SIFT and HOG while being more robust to distortions characteristic of large imagery datasets collected through crowdsourcing and citizen science.

Marketing campaigns are a set of strategic activities that can promote a business's goal. The effect prediction for marketing campaigns in a real industrial scenario is very complex and challenging due to the fact that prior knowledge is often learned from observation data, without any intervention for the marketing campaign. Furthermore, each subject is always under the interference of several marketing campaigns simultaneously. Therefore, we cannot easily parse and evaluate the effect of a single marketing campaign. To the best of our knowledge, there are currently no effective methodologies to solve such a problem, i.e., modeling an individual-level prediction task based on a hierarchical structure with multiple intertwined events. In this paper, we provide an in-depth analysis of the underlying parse tree-like structure involved in the effect prediction task and we further establish a Hierarchical Capsule Prediction Network (HapNet) for predicting the effects of marketing campaigns. Extensive results based on both the synthetic data and real data demonstrate the superiority of our model over the state-of-the-art methods and show remarkable practicability in real industrial applications.

Automated graph learning has drawn widespread research attention due to its great potential to reduce human efforts when dealing with graph data, among which hyperparameter optimization (HPO) is one of the mainstream directions and has made promising progress. However, how to obtain reliable and trustworthy prediction results with automated graph neural networks (GNN) is still quite underexplored. To this end, we investigate automated GNN calibration by marrying uncertainty estimation to the HPO problem. Specifically, we propose a hyperparameter uncertainty-induced graph convolutional network (HyperU-GCN) with a bilevel formulation, where the upper-level problem explicitly reasons uncertainties by developing a probabilistic hypernetworks through a variational Bayesian lens, while the lower-level problem learns how the GCN weights respond to a hyperparameter distribution. By squeezing model uncertainty into the hyperparameter space, the proposed HyperU-GCN could achieve calibrated predictions in a similar way to Bayesian model averaging over hyperparameters. Extensive experimental results on six public datasets were provided in terms of node classification accuracy and expected calibration error (ECE), demonstrating the effectiveness of our approach compared with several state-of-the-art uncertainty-aware and calibrated GCN methods.

Unsupervised domain adaptation (UDA) has become an appealing approach for knowledge transfer from a labeled source domain to an unlabeled target domain. However, when the classes in source and target domains are imbalanced, most existing UDA methods experience significant performance drop, as the decision boundary usually favors the majority classes. Some recent class-imbalanced domain adaptation (CDA) methods aim to tackle the challenge of biased label distribution by exploiting pseudo-labeled target samples during training process. However, these methods suffer from the issues with unreliable pseudo labels and error accumulation during training. In this paper, we propose a pairwise adversarial training approach for class-imbalanced domain adaptation. Unlike conventional adversarial training in which the adversarial samples are obtained from the lp ball of the original samples, we generate adversarial samples from the interpolated line of the aligned pairwise samples from source and target domains. The pairwise adversarial training (PAT) is a novel data-augmentation method which can be integrated into existing UDA models to tackle with the CDA problem. Experimental results and ablation studies show that the UDA models integrated with our method achieve considerable improvements on benchmarks compared with the original models as well as the state-of-the-art CDA methods. Our source code is available at: https://github.com/DamoSWL/Pairwise-Adversarial-Training

Multi-modal time series classification (MTC) uses complementary information from different modalities to improve the learning performance. Obtaining informative modality-specific representation plays an essential role in MTC. Attention mechanism has been widely adopted as an effective strategy for discovering discriminative cues underlying temporal data. However, most existing MTC methods only utilize attention to balance the feature weights within or cross modalities but ignore digging latent patterns from mutual-support information in attention space. Specifically, the attention distributions are different for multiple modalities which are supportive and instructional with each other. To this end, we propose a collaborative attention mechanism (CAM) for MTC based on a novel perspective to utilize attention module. CAM detects the attention differences among multi-modal time series, and adaptively integrates different attention information to benefit each other. We extend the long short-term memory (LSTM) to a Mutual-Aid RNN (MAR) for multi-modal collaboration. CAM takes advantages of modality-specific attention to guide another modality and discover potential information which is hard to be explored by itself. It paves a novel way of employing attention to enhance the capacity of multi-modal representations. Extensive experiments on four multi-modal time series datasets illustrate the CAM effectiveness to improve the single-modal and also boost multi-modal performances.

The foremost challenge to causal inference with real-world data is to handle the imbalance in the covariates with respect to different treatment options, caused by treatment selection bias. To address this issue, recent literature has explored domain-invariant representation learning based on different domain divergence metrics (e.g., Wasserstein distance, maximum mean discrepancy, position-dependent metric, and domain overlap). In this paper, we reveal the weaknesses of these strategies, i.e., they lead to the loss of predictive information when enforcing the domain invariance; and the treatment effect estimation performance is unstable, which heavily relies on the characteristics of the domain distributions and the choice of domain divergence metrics. Motivated by information theory, we propose to learn the Infomax and Domain-Independent Representations to solve the above puzzles. Our method utilizes the mutual information between the global feature representations and individual feature representations, and the mutual information between feature representations and treatment assignment predictions, in order to maximally capture the common predictive information for both treatment and control groups. Moreover, our method filters out the influence of instrumental and irrelevant variables, and thus it effectively increases the predictive ability of potential outcomes. Experimental results on both the synthetic and real-world datasets show that our method achieves state-of-the-art performance on causal effect inference. Moreover, our method exhibits reliable prediction performances when facing data with different characteristics of data distributions, complicated variable types, and severe covariate imbalance.

Unsupervised domain adaptation aims to learn domain-invariant features across domains to transfer knowledge from a well-labeled source domain to an unlabeled target domain. Recently, some unsupervised domain adaptation methods focus on semantically aligning data distributions with pseudo-labels of the target domain. However, semantic alignment based on pseudo-labels has potential risks, e.g., inaccurate pseudo-labeling from classifier, and error accumulation from pseudo-label bias. To alleviate these risks, we propose a novel self-supervision based semantic alignment (S3A) approach for unsupervised domain adaptation, which can jointly incorporate the source alignment and cross-domain target alignment for better semantic alignment across domains. S3A consists of a two-stage semantic alignment procedure with self-supervision. One is to capture the discriminative structure of source domain by aligning source data to source class prototypes, and the other is to match each target data to its neighbor in source domain with self-supervision. Extensive experiments on three benchmark datasets demonstrate the effectiveness of our proposed method, compared with the representative adversarial learning and self-supervised learning based unsupervised domain adaptation methods.

Search engines for domain-specific media collections often rely on rich metadata being available for the content items. The annotations may not be complete or rich enough to support an adequate retrieval effectiveness. As a result, some search queries receive only a small result set (low recall) and others might suffer from reduced relevance (low precision). To alleviate this, we present a framework that exploits external knowledge to provide entity-oriented reformulation suggestions for queries that contain entities. We propose that queries be added as surrogate nodes to an external Knowledge Graph (KG) via the use of state-of-the-art entity linking algorithms. Embedding methods are invoked on the augmented graph, which contains additional edges between surrogate nodes and KG entities. We introduce a new evaluation setting to evaluate the quality of these embeddings. Experimental results on seven datasets confirm the effectiveness of the approach.

Inferring causal effect from observational data has attracted much attention from various domains. Under the potential outcome framework, the estimation of counterfactuals is crucial for the investigation of causal effect at the individual level. Existing representation learning approaches focus on learning one balanced feature space, which ignores certain information predictive to the outcomes. To fully utilize the predictive information, we propose a Subspace learning based Counterfactual Inference (SCI) method to estimate causal effect at the individual level. Different from existing work, SCI learns both a common subspace, which preserves the information across all the treatment groups, and treatment-specific subspaces, which retain the information associated with each specific treatment. Learning from two kinds of subspaces helps SCI obtain better causal effect estimations than state-of-the-art methods, demonstrated by a series of experiments on synthetic and real-world datasets.

This paper proposes a new physics-guided machine learning approach that incorporates the scientific knowledge in physics-based models into machine learning models. Physics-based models are widely used to study dynamical systems in a variety of scientific and engineering problems. Although they are built based on general physical laws that govern the relations from input to output variables, these models often produce biased simulations due to inaccurate parameterizations or approximations used to represent the true physics. In this paper, we aim to build a new data-driven framework to monitor dynamical systems by extracting general scientific knowledge embodied in simulation data generated by the physics-based model. To handle the bias in simulation data caused by imperfect parameterization, we propose to extract general physical relations jointly from multiple sets of simulations generated by a physics-based model under different physical parameters. In particular, we develop a spatio-temporal network architecture that uses its gating variables to capture the variation of physical parameters. We initialize this model using a pre-training strategy that helps discover common physical patterns shared by different sets of simulation data. Then we fine-tune it using limited observation data via a contrastive learning process. By leveraging the complementary strength of machine learning and domain knowledge, our method has been shown to produce accurate predictions, use less training samples and generalize to out-of-sample scenarios. We further show that the method can provide insights about the variation of physical parameters over space and time in two domain applications: predicting temperature in streams and predicting temperature in lakes.

Universal domain adaptation (UniDA) aims to transfer knowledge from a labeled source domain to an unlabeled target domain where both domains share a common label space and hold a private label space respectively. One of the most challenging goals in UniDA is to separate target samples from common classes and these from private classes without any prior knowledge on the target label space. In this paper, we propose a novel self-supervised adaptive memory network with consistency regularization for UniDA. The adaptive memory includes all target samples and source class centers, which dynamically divides target samples into common area, uncertain area, and unknown area based on the entropy. Our proposed framework jointly assigns a specific neighborhood to each target sample and clusters the target sample to its neighbor from the neighborhood. Most importantly, the proposed framework adopts consistency regularization that gradually makes the output of the classifier more reliable. This simple strategy is proved to be very effective for UniDA problem. Experimental results on two UniDA benchmarks demonstrate the effectiveness of our method.

Treatment effect estimation from observational data is a critical research topic across many domains. The foremost challenge in treatment effect estimation is how to capture hidden confounders. Recently, the growing availability of networked observational data offers a new opportunity to deal with the issue of hidden confounders. Unlike networked data in traditional graph learning tasks, such as node classification and link detection, the networked data under the causal inference problem has its particularity, i.e., imbalanced network structure. In this paper, we propose a Graph Infomax Adversarial Learning (GIAL) model for treatment effect estimation, which makes full use of the network structure to capture more information by recognizing the imbalance in network structure. We evaluate the performance of our GIAL model on two benchmark datasets, and the results demonstrate superiority over the state-of-the-art methods.

In this paper, we propose a 2.5D pose guided human image generation method that integrates depth information with 2D poses. Given a target 2.5D pose and an image of a person, our method generates a new image of that person with the target pose. To incorporate depth information into the pose structure, we design a three-layer pose space that allows accurate pose transfer compared with regular 2D pose structure. Specifically, our pose space enables the generative models to address the occlusion problems commonly happened in human image generation and also helps recognize spatial front-back relations of limbs. Extensive quantitative and qualitative results on the DeepFashion and Human 3.6M datasets demonstrate the effectiveness of our method.

Owing to the remarkable capability of extracting effective graph embeddings, graph convolutional network (GCN) and its variants have been successfully applied to a broad range of tasks, such as node classification, link prediction, and graph classification. Traditional GCN models suffer from the issues of overfitting and oversmoothing, while some recent techniques like DropEdge could alleviate these issues and thus enable the development of deep GCN. However, training GCN models is non-trivial, as it is sensitive to the choice of hyperparameters such as dropout rate and learning weight decay, especially for deep GCN models. In this paper, we aim to automate the training of GCN models through hyperparameter optimization. To be specific, we propose a self-tuning GCN approach with an alternate training algorithm, and further extend our approach by incorporating the population based training scheme. Experimental results on three benchmark datasets demonstrate the effectiveness of our approaches on optimizing multi-layer GCN, compared with several representative baselines.

Causal inference is a critical research topic across many domains, such as statistics, computer science, education, public policy, and economics, for decades. Nowadays, estimating causal effect from observational data has become an appealing research direction owing to the large amount of available data and low budget requirement, compared with randomized controlled trials. Embraced with the rapidly developed machine learning area, various causal effect estimation methods for observational data have sprung up. In this survey, we provide a comprehensive review of causal inference methods under the potential outcome framework, one of the well-known causal inference frameworks. The methods are divided into two categories depending on whether they require all three assumptions of the potential outcome framework or not. For each category, both the traditional statistical methods and the recent machine learning enhanced methods are discussed and compared. The plausible applications of these methods are also presented, including the applications in advertising, recommendation, medicine, and so on. Moreover, the commonly used benchmark datasets as well as the open-source codes are also summarized, which facilitate researchers and practitioners to explore, evaluate and apply the causal inference methods.

Concept Factorization (CF) and its variants may produce inaccurate representation and clustering results due to the sensitivity to noise, hard constraint on the reconstruction error, and pre-obtained approximate similarities. To improve the representation ability, a novel unsupervised Robust Flexible Auto-weighted Local-coordinate Concept Factorization (RFA-LCF) framework is proposed for clustering high-dimensional data. Specifically, RFA-LCF integrates the robust flexible CF by clean data space recovery, robust sparse local-coordinate coding, and adaptive weighting into a unified model. RFA-LCF improves the representations by enhancing the robustness of CF to noise and errors, providing a flexible constraint on the reconstruction error and optimizing the locality jointly. For robust learning, RFA-LCF clearly learns a sparse projection to recover the underlying clean data space, and then the flexible CF is performed in the projected feature space. RFA-LCF also uses a L2,1-norm based flexible residue to encode the mismatch between the recovered data and its reconstruction, and uses the robust sparse local-coordinate coding to represent data using a few nearby basis concepts. For auto-weighting, RFA-LCF jointly preserves the manifold structures in the basis concept space and new coordinate space in an adaptive manner by minimizing the reconstruction errors on clean data, anchor points and coordinates. By updating the local-coordinate preserving data, basis concepts and new coordinates alternately, the representation abilities can be potentially improved. Extensive results on public databases show that RFA-LCF delivers the state-of-the-art clustering results compared with other related methods.

The dramatically growing availability of observational data is being witnessed in various domains of science and technology, which facilitates the study of causal inference. However, estimating treatment effects from observational data is faced with two major challenges, missing counterfactual outcomes and treatment selection bias. Matching methods are among the mostly widely used and fundamental approaches to estimating treatment effects, but existing matching methods have poor performance when facing data with high dimensional and complicated variables. We propose a feature selection representation matching (FSRM) method based on deep representation learning and matching, which maps the original covariate space into a selective, nonlinear, and balanced representation space, and then conducts matching in the learned representation space. FSRM adopts deep feature selection to minimize the influence of irrelevant variables for estimating treatment effects and incorporates a regularizer based on the Wasserstein distance to learn balanced representations. We evaluate the performance of our FSRM method on three datasets, and the results demonstrate superiority over the state-of-the-art methods.

Temporal data are continuously collected in a wide range of domains. The increasing availability of such data has led to significant developments of time series analysis. Time series classification, as an essential task in time series analysis, aims to assign a set of temporal sequences to different categories. Among various approaches for time series classification, the distance metric learning based ones, such as the virtual sequence metric learning (VSML), have attracted increased attention due to their remarkable performance. In VSML, virtual sequences attract samples from different classes to facilitate time series classification. However, the existing VSML methods simply employ fixed virtual sequences, which might not be optimal for the subsequent classification tasks. To address this issue, in this paper, we propose a novel time series classification method named Discriminative Virtual Sequence Learning (DVSL). Following the unified framework of sequence metric learning, our DVSL method jointly learns a set of discriminative virtual sequences that help separate time series samples in a feature space, and optimizes the temporal alignment by dynamic time warping. Extensive experiments on 15 UCR time series datasets demonstrate the efficiency of DVSL, compared with several representative baselines.

Causal inference has numerous real-world applications in many domains such as health care, marketing, political science and online advertising. Treatment effect estimation, a fundamental problem in causal inference, has been extensively studied in statistics for decades. However, traditional treatment effect estimation methods may not well handle large-scale and high-dimensional heterogeneous data. In recent years, an emerging research direction has attracted increasing attention in the broad artificial intelligence field, which combines the advantages of traditional treatment effect estimation approaches (e.g., matching estimators) and advanced representation learning approaches (e.g., deep neural networks). In this tutorial, we will introduce both traditional and state-of-the-art representation learning algorithms for treatment effect estimation. Background about causal inference, counterfactuals and matching estimators will be covered as well. We will also showcase promising applications of these methods in different application domains.

Online news recommender systems aim to make personalized recommendations according to user preferences, which require modeling users’ short-term reading interest. However, due to the limited logged user interactions in practice, news recommendation at session-level becomes very challenging. Existing methods on session-based news recommendation mainly focus on extracting features from news articles and sequential user-item interactions, but they usually ignore the semantic-level structural information among news articles and do not explore external knowledge sources. In this paper, we propose a novel Context-Aware Graph Embedding (CAGE) framework for session-based news recommendation, which builds an auxiliary knowledge graph to enrich the semantic meaning of entities involved in articles, and further refines the article embeddings by graph convolutional networks. Experimental results on a real-world news dataset demonstrate the effectiveness of our method compared with the state-of-the-art methods on session-based news recommendation.

In the past decade, we have envisioned enormous growth in the data generated by different sources, ranging from weather sensors and customer purchasing records to Internet of Things devices. Emerging data-driven technologies have been reforming our daily life for years, such as Amazon Personalize [1], which creates real-time individualized recommendations for customers according to multidimensional data analytics. It is, however, a challenging task to fully utilize and harness the potential of data, especially big data, due to Volume, Velocity, Variety, Variability and Value (5Vs) [2]. Most businesses thus choose to migrate their hardware demands to cloud providers, such as Amazon Web Service [3], which is powered by hundreds of thousands of servers. A cluster that builds up by a number of cloud servers is a basic management unit to provide shared computing resources. The typical structure of a cluster consists of managers and workers. When a job arrives at the cluster, as the first step, managers have to select a worker to host the incoming job. Traditionally, the selection process is based on the state of the workers, e.g., resource availability and specifications of jobs, e.g., labels, zones and regions. With respect to currently running jobs, we propose a progress based container placement scheme, named ProCon. When scheduling incoming containers, ProCon not only considers instant resource utilization on the workers but also takes into account the estimation of future resource usage. Through monitoring the progress of running jobs, ProCon balances the resource contentions across the cluster and reduces the completion time as well as the makespan. Specifically, extensive experiments prove that ProCon reduces completion time by up to 53.3% for a particular job and improves overall performance by 23.0%. Additionally, ProCon records an improvement of makespan for up to 37.4% when compared to the default scheduler available in Kubernetes.

In this paper, we propose a robust representation learning model called Adaptive Structure-constrained Low-Rank Coding (AS-LRC) for the latent representation of data. To recover the underlying subspaces more accurately, AS-LRC seamlessly integrates an adaptive weighting based block-diagonal structure-constrained low-rank representation and the group sparse salient feature extraction into a unified framework. Specifically, AS-LRC performs the latent decomposition of given data into a low-rank reconstruction by a block-diagonal codes matrix, a group sparse locality-adaptive salient feature part and a sparse error part. To enforce the block-diagonal structures adaptive to different real datasets for the low-rank recovery, AS-LRC clearly computes an auto-weighting matrix based on the locality-adaptive features and multiplies by the low-rank coefficients for direct minimization at the same time. This encourages the codes to be block-diagonal and can avoid the tricky issue of choosing optimal neighborhood size or kernel width for the weight assignment, suffered in most local geometrical structures-preserving low-rank coding methods. In addition, our AS-LRC selects the L2, 1-norm on the projection for extracting group sparse features rather than learning low-rank features by Nuclear-norm regularization, which can make learnt features robust to noise and outliers in samples, and can also make the feature coding process efficient. Extensive visualizations and numerical results demonstrate the effectiveness of our AS-LRC for image representation and recovery.

Treatment effect estimation refers to the estimation of causal effects, which benefits decision-making process across various domains, but it is a challenging problem in real practice. The estimation of causal effects from observational data at the individual level faces two major challenges, i.e., treatment selection bias and missing counterfactuals. Existing methods tackle the selection bias problem by learning a balanced representation and infer the missing counterfactuals based on the learned representation. However, most existing methods learn the representation in a global manner and ignore the local similarity information, which is essential for an accurate estimation of causal effects. Motivated by the above observations, we propose a novel representation learning method, which adaptively extracts fine-grained similarity information from the original feature space and minimizes the distance between different treatment groups as well as the similarity loss during the representation learning procedure. Experiments on three public datasets demonstrate that the proposed method achieves the best performance in causal effect estimation among all the compared methods and is robust to the treatment selection bias.

The last decade has witnessed a surge of interest in applying deep learning models for discovering sequential patterns from a large volume of data. Recent works show that deep learning models can be further improved by enforcing models to learn a smooth output distribution around each data point. This can be achieved by augmenting training data with slight perturbations that are designed to alter model outputs. Such adversarial training approaches have shown much success in improving the generalization performance of deep learning models on static data, e.g., transaction data or image data captured on a single snapshot. However, when applied to sequential data, the standard adversarial training approaches cannot fully capture the discriminative structure of a sequence. This is because real-world sequential data are often collected over a long period of time and may include much irrelevant information to the classification task. To this end, we develop a novel adversarial training approach for sequential data classification by investigating when and how to perturb a sequence for an effective data augmentation. Finally, we demonstrate the superiority of the proposed method over baselines in a diversity of real-world sequential datasets.

Modeling user behavior from unstructured software log-trace data is critical in providing personalized service (\emphe.g., cross-platform recommendation). Existing user modeling approaches cannot well handle the long-term temporal information in log data, or produce semantically meaningful results for interpreting user logs. To address these challenges, we propose a Log2Intent framework for interpretable user modeling in this paper. Log2Intent adopts a deep sequential modeling framework that contains a temporal encoder, a semantic encoder and a log action decoder, and it fully captures the long-term temporal information in user sessions. Moreover, to bridge the semantic gap between log-trace data and human language, a recurrent semantics memory unit (RSMU) is proposed to encode the annotation sentences from an auxiliary software tutorial dataset, and the output of RSMU is fed into the semantic encoder of Log2Intent. Comprehensive experiments on a real-world Photoshop log-trace dataset with an auxiliary Photoshop tutorial dataset demonstrate the effectiveness of the proposed Log2Intent framework over the state-of-the-art log-trace user modeling method in three different tasks, including log annotation retrieval, user interest detection and user next action prediction.

Detecting land use and land cover changes is critical to monitor natural resources and analyze global environmental changes. In this paper, we investigate the land cover detection using the remote sensing data from earth-observing satellites. Due to the natural disturbances, e.g., clouds and aerosoles, and the data acquisition errors by devices, remote sensing data frequently contain much noise. Also, many land covers cannot be easily identified in most dates of a year. Instead, they show distinctive temporal patterns only during certain period of a year, which is also referred to as the discriminative period. To address these challenges, we propose a novel framework which combines the spatial context knowledge with the LSTM-based temporal modeling for land cover detection. Specifically, the framework learns the spatial context knowledge selectively from its neighboring locations. Then we propose two approaches for discriminative period detection based on multi-instance learning and local attention mechanism, respectively. Our evaluations in two real-world applications demonstrate the effectiveness of the proposed method in identifying land covers and detecting discriminative periods.

Factorization Machines (FM) have been widely applied in industrial applications for recommendations. Traditionally FM models are trained in batch mode, which entails training the model with large datasets every few hours or days. Such training procedure cannot capture the trends evolving in real time with large volume of streaming data. In this paper, we propose an online training scheme for FM with the alternating least squares (ALS) technique, which has comparable performance with existing batch training algorithms. We incorporate an online update mechanism to the model parameters at the cost of storing a small cache. The mechanism also stabilizes the training error more than a traditional online training technique like stochastic gradient descent (SGD) as data points come in, which is crucial for real-time applications. Experiments on large scale datasets validate the efficiency and robustness of our method.

Understanding user behavior and predicting future behavior on the web is critical for providing seamless user experiences as well as increasing revenue of service providers. Recently, thanks to the remarkable success of recurrent neural networks (RNNs), it has been widely used for modeling sequences of user behaviors. However, although sequential behaviors appear across multiple domains in practice, existing RNN-based approaches still focus on the single-domain scenario assuming that sequential behaviors come from only a single domain. Hence, in order to analyze sequential behaviors across multiple domains, they require to separately train multiple RNN models, which fails to jointly model the interplay among sequential behaviors across multiple domains. Consequently, they often suffer from lack of information within each domain. In this paper, we first introduce a practical but overlooked phenomenon in sequential behaviors across multiple domains, i.e.,domain switch where two successive behaviors belong to different domains. Then, we propose aDomain Switch-Aware Holistic Recurrent Neural Network (DS-HRNN) that effectively shares the knowledge extracted from multiple domains by systematically handlingdomain switch for the multi-domain scenario. DS-HRNN jointly models the multi-domain sequential behaviors and accurately predicts the future behaviors in each domain with only a single RNN model. Our extensive evaluations on two real-world datasets demonstrate that \DCHRNN\ outperforms existing RNN-based approaches and non-sequential baselines with significant improvements by up to 14.93% in terms of recall of the future behavior prediction.

Ensemble Clustering (EC) is an important topic for data cluster analysis. It targets to integrate multiple Basic Partitions (BPs) of a particular dataset into a consensus partition. Among previous works, one promising and effective way is to transform EC as a graph partitioning problem on the co-association matrix, which is a pair-wise similarity matrix summarized by all the BPs in essence. However, most existing EC methods directly utilize the co-association matrix, yet without considering various noises (e.g., the disagreement between different BPs and the outliers) that may exist in it. These noises can impair the cluster structure of a co-association matrix, and thus mislead the final graph partitioning process. To address this challenge, we propose a novel Robust Spectral Ensemble Clustering (RSEC) algorithm in this article. Specifically, we learn low-rank representation (LRR) for the co-association matrix to uncover its cluster structure and handle the noises, and meanwhile, we perform spectral clustering with the learned representation to seek for a consensus partition. These two steps are jointly proceeded within a unified optimization framework. In particular, during the optimizing process, we leverage consensus partition to iteratively enhance the block-diagonal structure of LRR, in order to assist the graph partitioning. To solve RSEC, we first formulate it by using nuclear norm as a convex proxy to the rank function. Then, motivated by the recent advances in non-convex rank minimization, we further develop a non-convex model for RSEC and provide it a solution by the majorization--minimization Augmented Lagrange Multiplier algorithm. Experiments on 18 real-world datasets demonstrate the effectiveness of our algorithm compared with state-of-the-art methods. Moreover, several impact factors on the clustering performance of our approach are also explored extensively.

In this paper, we propose a novel unsupervised Nonnegative Adaptive Feature Extraction (NAFE) algorithm for data representation and classification. The formulation of NAFE integrates the sparsity constrained nonnegative matrix factorization (NMF), representation learning, and adaptive reconstruction weight learning into a unified model. Specifically, NAFE performs feature and weight learning over the new robust representations of NMF for more accurate measure and representation. For nonnegative adaptive feature extraction, our NAFE first utilizes the sparsity constrained NMF to obtain the new and robust representations of the original data. To preserve the manifold structures of the learnt new representations, we also incorporate a neighborhood reconstruction error over the weight matrix for joint minimization. Note that to further improve the representation power, the weights are jointly shared in the new low-dimensional nonnegative representation space, low-dimensional nonlinear manifold space, and low-dimensional projective subspace, i.e., local neighborhood information is clearly preserved in different feature spaces so that informative representations and features can be jointly obtained. To enable NAFE to extract features from new data, we also include a feature approximation error by a linear projection so that the learnt extractor can obtain features from new data efficiently. Extensive simulations show that our formulation can deliver state-of-the-art results on several public databases for feature extraction and classification, compared with several related methods.

Detecting outliers or anomalies is a fundamental problem in various machine learning and data mining applications. Conventional outlier detection algorithms are mainly designed for single-view data. Nowadays, data can be easily collected from multiple views, and many learning tasks such as clustering and classification have benefited from multi-view data. However, outlier detection from multi-view data is still a very challenging problem, as the data in multiple views usually have more complicated distributions and exhibit inconsistent behaviors. To address this problem, we propose a multi-view low-rank analysis (MLRA) framework for outlier detection in this article. MLRA pursuits outliers from a new perspective, robust data representation. It contains two major components. First, the cross-view low-rank coding is performed to reveal the intrinsic structures of data. In particular, we formulate a regularized rank-minimization problem, which is solved by an efficient optimization algorithm. Second, the outliers are identified through an outlier score estimation procedure. Different from the existing multi-view outlier detection methods, MLRA is able to detect two different types of outliers from multiple views simultaneously. To this end, we design a criterion to estimate the outlier scores by analyzing the obtained representation coefficients. Moreover, we extend MLRA to tackle the multi-view group outlier detection problem. Extensive evaluations on seven UCI datasets, the MovieLens, the USPS-MNIST, and the WebKB datasets demon strate that our approach outperforms several state-of-the-art outlier detection methods.

The lack of labels and the poor quality of data present a common challenge in many data mining and machine learning problems. The model performance might be limited if only a few labeled samples are available for training. Moreover, the data may be noisy in reality, which disturbs the data distribution and further hinders the learning performance. These problems become even more critical in multi-label classification, which has an intricate label space and usually requires clean data for training. In this paper, we aim to tackle the above problems by learning effective feature representations for semi-supervised multi-label classification. We propose a novel approach named Adaptive Low-rank Semi-supervised learning for Multi-label classification (ALSM). It learns an intermediate feature space for both labeled and unlabeled training samples via low-rank matrix recovery, and employs an adaptive semi-supervised learning strategy to train a multi-label classifier. We solve the problem by devising an optimization algorithm based on the Alternating Direction Method of Multipliers (ADMM). Our approach can be applied to both transductive and inductive semi-supervised multi-label classification problems. Experiments on five benchmark multi-label datasets show that our approach outperforms the representative multi-label classification methods in most cases.

Graphs play an essential role in many data mining paradigms, such as semi-supervised classification. Conventional graph learning methods mainly focus on constructing graphs from single-view data. Nowadays data can be collected from multiple views using various sensors. How to construct a robust and reliable graph from multi-view data is still an open problem. In this paper, we propose a multi-view graph learning (MVGL) approach with adaptive label propagation for semi-supervised classification. MVGL integrates latent factor extraction, graph sparsification, and label propagation into a unified framework. It seeks shared latent factors from multi-view data as view-independent data representations, and then constructs a sparse graph accordingly. Meanwhile, the label propagation is adaptively optimized during graph construction. An efficient optimization algorithm is designed to solve the model. Experimental results on two benchmark datasets show remarkable improvements over both single-view and multi-view learning baselines.

By virtue of the increasingly large amount of various sensors, information about the same object can be collected from multiple views. These mutually enriched information can help many real-world applications, such as daily activity recognition in which both video cameras and on-body sensors are continuously collecting information. Such multivariate time series (m.t.s.) data from multiple views can lead to a significant improvement of classification tasks. However, the existing methods for time series data classification only focus on single-view data, and the benefits of mutual-support multiple views are not taken into account. In light of this challenge, we propose a novel approach, named Multi-view Discriminative Bilinear Projections (MDBP), for extracting discriminative features from multi-view m.t.s. data. First, MDBP keeps the original temporal structure of m.t.s. data, and projects m.t.s. from different views onto a shared latent subspace. Second, MDBP incorporates discriminative information by minimizing the within-class separability and maximizing the between-class separability of m.t.s. in the shared latent subspace. Moreover, a Laplacian regularization term is designed to preserve the temporal smoothness within m.t.s.. Extensive experiments on two real-world datasets demonstrate the effectiveness of our approach. Compared to the state-of-the-art multi-view learning and m.t.s. classification methods, our approach greatly improves the classification accuracy due to the full exploration of multi-view streaming data. Moreover, by using a feature fusion strategy, our approach further improves the classification accuracy by at least 10%.

Ensemble Clustering (EC) aims to integrate multiple Basic Partitions (BPs) of the same dataset into a consensus one. It could be transformed as a graph partition problem on the co-association matrix derived from BPs. However, existing EC methods usually directly use the co-association matrix, yet without considering various noises (e.g., the disagreement between different BPs or outliers) that may exist in it. These noises can impair the cluster structure of a co-association matrix and thus degrade the final clustering performance. In this paper, we propose a novel Robust Spectral Ensemble Clustering (RSEC) approach to address this challenge. First, RSEC learns a robust representation for the co-association matrix through low-rank constraint, which reveals the cluster structure of a co-association matrix and captures various noises in it. Second, RSEC finds the consensus partition by conducting spectral clustering. These two steps are iteratively performed in a unified optimization framework. Most importantly, during our optimization process, we utilize consensus partition to iteratively enhance the block-diagonal structure of the learned representation to further assist the clustering process. Experiments on numerous real-world datasets demonstrate the effectiveness of our method compared with the state-of-the-art. Moreover, several impact factors that may affect the clustering performance of our approach are also explored extensively.

Image clustering has been a critical preprocessing step for vision tasks, e.g., visual concept discovery, content-based image retrieval. Conventional image clustering methods use handcraft visual descriptors as basic features via K-means, or build the graph within spectral clustering. Recently, representation learning with deep structure shows appealing performance in unsupervised feature pre-treatment. However, few studies have discussed how to deploy deep representation learning to image clustering problems, especially the unified framework which integrates both representation learning and ensemble clustering for efficient image clustering still remains void. In addition, even though it is widely recognized that with the increasing number of basic partitions, ensemble clustering gets better performance and lower variances, the best number of basic partitions for a given data set is a pending problem. In light of this, we propose the Infinite Ensemble Clustering (IEC), which incorporates the power of deep representation and ensemble clustering in a one-step framework to fuse infinite basic partitions. Generally speaking, a set of basic partitions is firstly generated from the image data, then by converting the basic partitions to the 1-of-K codings, we link the marginalized auto-encoder to the infinite ensemble clustering with i.i.d. basic partitions, which can be approached by the closed-form solutions, finally we follow the layer-wise training procedure and feed the concatenated deep features to K-means for final clustering. Extensive experiments on diverse vision data sets with different levels of visual descriptors demonstrate both the time efficiency and superior performance of IEC compared to the state-of-the-art ensemble clustering and deep clustering methods.

Collaborative filtering (CF) has been widely employed within recommender systems to solve many real-world problems. Learning effective latent factors plays the most important role in collaborative filtering. Traditional CF methods based upon matrix factorization techniques learn the latent factors from the user-item ratings and suffer from the cold start problem as well as the sparsity problem. Some improved CF methods enrich the priors on the latent factors by incorporating side information as regularization. However, the learned latent factors may not be very effective due to the sparse nature of the ratings and the side information. To tackle this problem, we learn effective latent representations via deep learning. Deep learning models have emerged as very appealing in learning effective representations in many applications. In particular, we propose a general deep architecture for CF by integrating matrix factorization with deep feature learning. We provide a natural instantiations of our architecture by combining probabilistic matrix factorization with marginalized denoising stacked auto-encoders. The combined framework leads to a parsimonious fit over the latent features as indicated by its improved performance in comparison to prior state-of-art models over four large datasets for the tasks of movie/book recommendation and response prediction.

Outlier detection is a fundamental problem in data mining. Unlike most existing methods that are designed for single-view data, we propose a multi-view outlier detection approach in this paper. Multi-view data can provide plentiful information of samples, however, detecting outliers from multi-view data is still a challenging problem due to the complicated distribution and inconsistent behavior of samples across different views. We address this problem through robust data representation, by building a Multi-view Low-Rank Analysis (MLRA) framework. Our framework contains two major components. First, it performs cross-view low-rank analysis for revealing the intrinsic structures of data. Second, it identifies outliers by estimating the outlier score for each test sample. Specifically, we formulate the cross-view low-rank analysis as a constrained rank-minimization problem, and present an efficient optimization algorithm to solve it. Different from the existing multi-view outlier detection methods, our framework is able to detect two different types of outliers from multiple views simultaneously. To this end, we design a criterion to estimate the outlier scores by analyzing the obtained representation coefficients. Experimental results on seven UCI datasets and the USPS-MNIST dataset demonstrate that our approach outperforms several state-of-the-art single-view and multi-view outlier detection methods in most cases.

Conversion prediction and click prediction are two important and intertwined problems in display advertising, but existing approaches usually look at them in isolation. In this paper, we aim to predict the conversion response of users by jointly examining the past purchase behavior and the click response behavior. Additionally, we model the temporal dynamics between the click response and purchase activity into a unified framework. In particular, a novel matrix factorization approach named dynamic collective matrix factorization (DCMF) is proposed to address this problem. Our model considers temporal dynamics of post-click conversions and also takes advantages of the side information of users, advertisements, and items. Experiments on a real-world marketing dataset show that our model achieves significant improvements over several baselines.

Graphs have been widely applied in modeling the relationships and structures in real-world applications. Graph construction is the most critical part in these models, while how to construct an effective graph is still an open problem. In this paper, we propose a novel approach to graph construction based on two observations. First, by virtue of recent advances in low-rank subspace recovery, the similarity between every two samples evaluated in the low-rank code space is more robust than that in the sample space. Second, a sparse and balanced graph can greatly increase the performance of learning tasks, such as label propagation in graph based semi-supervised learning. The$k$-NN sparsification can provide fast solutions to constructing unbalanced sparse graphs, and$b$-matching constraint is a necessary route for generating balanced graphs. These observations motivate us to jointly learn the low-rank codes and balanced (or unbalanced) graph simultaneously. In particular, two non-convex models are built by incorporating$k$-NN constraint and$b$-matching constraint into the low-rank representation model, respectively. We design a majorization-minimization augmented Lagrange multiplier (MM-ALM) algorithm to solve the proposed models. Extensive experimental results on four image databases demonstrate the superiority of our graphs over several state-of-the-art graphs in data clustering, transductive and inductive semi-supervised learning.

This work focuses on early classification of ongoing observation of the object, which is beneficial for a number of applications that require time-critical decision making. We propose an approach for discovering two key aspects of multivariate time series (m.t.s.) observation, (1) Temporal Dynamics and (2) Sequential Cues. The key idea is that m.t.s. Observation can be represented as an instantiation of a Multivariate Marked Point-Process (Multi-MPP). Each variable characterizes the temporal dynamics of a particular feature event of an object, where both timing and strength information of that feature event are preserved. To make this model computationally practical, we introduce the Multilevel-Discretized Marked Point-Process (MD-MPP) model which can ensure a good piece-wise stationary property both in the time-domain and mark-space while preserving dynamics as much as possible. Based on this model, another important temporal patterns of early classification, sequential cues among variables, becomes formalizable. We construct a probabilistic suffix tree to represent sequential patterns among features in terms of Variable order Markov Model (VMM). The effectiveness of our approach is evaluated on three experimental scenarios. Our method achieves superior performance for early classification of ongoing m.t.s. Observation data.

Subspace learning is a popular approach for feature extraction and classification. However, its performance would be heavily degraded when data are corrupted by large amounts of noise. Inspired by recent work in matrix recovery, we tackle this problem by exploiting a subspace that is robust to noise and large variability for classification. Specifically, we propose a novel Supervised Regularization based Robust Subspace (SRRS) approach via low-rank learning. Unlike existing subspace methods, our approach jointly learns low-rank representations and a robust subspace from noisy observations. At the same time, to improve the classification performance, class label information is incorporated as supervised regularization. The problem can then be formulated as a constrained rank minimization objective function, which can be effectively solved by the inexact augmented Lagrange multiplier (ALM) algorithm. Our approach differs from current sparse representation and low-rank learning methods in that it explicitly learns a low-dimensional subspace where the supervised information is incorporated. Extensive experimental results on four datasets demonstrate that our approach outperforms the state-of-the-art subspace and low-rank learning methods in almost all cases, especially when the data contain large variations or are heavily corrupted by noise.