Inductive knowledge graph completion (KGC) aims to represent unseen entities and complete triplets in emerging knowledge graphs (KGs), while the existing studies ignore that unseen elements combined with seen ones constitute a holistic new relational graph, where emerging KGs have inescapable impacts backtracking to original ones. Therefore, it is not only necessary to predict triplets in emerging KGs, but also with particular significance to further improve the completeness of original ones, considering the semantic and topological variations in the holistic new graph. To fill in this gap, we formulate a newIT(Inductive-Transductive) KGC task to transductively complete triplets inside original KGs after entities in the emerging scenario are represented and fine-tuned in an inductive manner. In order to handle this task, a novel model entitled StaR (Self-adaptive Retroaction-awareRepresentation) is proposed consisting of the following two modules: 1) a self-adaptive semantic encoding network is designed to adaptively adjust embeddings of seen entities to their surrounding semantic mutations; 2) a relation-aware transformer layer is developed to represent both seen and unseen entities in a unified representation space and generalize evolving reasoning paradigms to the whole graph. Our experimental results demonstrate that, compared with state-of-the-art methods, StaR is not only competitive in inductive KGC for unseen entities, but also ulteriorly improves the completeness of original parts inside the holistic new relational graph in ourITKGC task.

Deep Reinforcement Learning (DRL) has garnered significant attention as a promising approach for developing intelligent and adaptive recommender systems. This paradigm is especially well-suited for recommendation scenarios characterized by dynamic user environments, temporally delayed feedback signals, and continuously shifting user preferences. However, deploying DRL within recommendation scenarios introduces a range of intricate challenges. These include the design of meaningful and task-aligned reward functions, effective navigation of vast and complex action spaces, and the need to maintain sample efficiency in data-sparse environments. Ensuring robust and stable training dynamics adds further difficulty. This special issue brings together a diverse collection of cutting-edge research that addresses these pressing challenges, showcasing advances that move the field toward more adaptive, robust, and personalized recommendation systems grounded in reinforcement learning.

The information available in multi-attributed road-social networks includes network structure, location information, and numerical attributes. Most studies mainly focus on mining communities by combining structure with attributes or structure with location, which do not consider structure, attributes, and location simultaneously. Therefore, we propose a parameter-free algorithm, called LCDMRS, to mine local communities in multi-attributed road-social networks. LCDMRS extracts a sub-network surrounding the given node and embeds it to generate the vector representations of nodes, which incorporates both structural and attributed information. Based on the vector representations of nodes, the average cosine similarity between nodes is designed to ensure both the structural and attributed cohesiveness of the community, while the community node density is designed to ensure the spatial cohesiveness of the community. Targeting the community node density and cosine similarity of nodes, LCDMRS takes the given node as the starting node and employs the community dominance relation to expand the community outward. Experimental results on multiple real-world datasets demonstrate LCDMRS outperforms comparison algorithms.

Nowadays, detecting sentiment or emotion from user generated texts has been intensively studied in natural language understanding, especially via neural-based models based on text representation. However, the interpretability on how could the final text sentiment be determined by neural-based text representation has not been thoroughly unfolded yet. Consequently, in this paper, we proposeCogLignwhich injects theneural-cognitionderived from Electroencephalogram (EEG)-signal into theneural-basedtext sentiment analysis model, aimed at learning the activation of brain regions stimulated by different sentiments, so as to guide our proposedCogLignto make proper determination on text sentiment in brain-like way. Specifically, on the one hand, the given videos in different sentiments have been watched bysubjects, during which the EEG-signals are monitored to construct brain connectivity pattern asbrain graph(BG), attaining more obvious sentiment response on brain region activation forneural-cognition. On the other hand, we interpret the video-plots (or video-semantics) along timeline into text, where the entire video-interpreted-text will bestrictly boundwith the wholeEEG-signal-sequencebysegmentvia the fixed size oftime-window. Then, entities and relations are extracted from the video-interpreted-text to constructknowledge graph(KG), depicting text semantics. Next, mapping fromentities(or nodes) inKGtoEEG-Electrodes(or nodes) inBG, further dated back to different brain regions, has been learned viacognition alignmentbetween the EEG-derivedBGand text-derivedKG. In this way, by aligningneural cognitionfrombrain graphwith thesemantic cognitionfromknowledge graph, our proposed frameworkCogLigncan not only achieve the overall best sentiment analysis performance on thevideo-interpreted-text, but can also detect brain connectivity patterns in different sentiments more consistent with the prior conclusion of brain region sentiment preference, revealing competitiveinterpretabilityon text sentiment determination.

The tremendous positive driving effect of Graph Convolutional Network (GCN) and Graph Contrastive Learning (GCL) for recommender systems has become a consensus. GCN encoders are extensively used in recommendation models for capturing high-order connectivities between users and items, whereas GCL accelerates the training of recommendation tasks by adding extra supervision signals from contrastive objectives. However, little attention has been paid on corresponding theories that are truly tailored to recommendation tasks. From the technical perspective, Collaborative Filtering (CF) is seen as an important factor in recommender systems. It is applied to measure user–user, item–item, and user–item similarities rather than to achieve better clustering or node classification results. Besides, heuristic-based data augmentation may not be hold true in the field of recommender systems as it requires additional training costs and introduces noises that will corrupt the interaction graph structure and the semantic information of nodes. To tackle these limitations, we propose a novel Embedding-Less Graph Collaborative Filtering (EGCF) for recommendation, which is tailor-made for the problem mentioned for CF and further simplifies existing solutions. Structurally, it consists of two parts: embedding-less GCN and embedding-less GCL. The former improves user–item affinity by streamlining user-type embeddings and carrying out iterative graph convolution. And the latter utilizes three-type contrastive objectives to directly measure the alignment and the uniformity of users, items, and interaction pairs, respectively, avoiding any type of data augmentation or multi-view construction. Even though EGCF has been extremely streamlined, extensive experimental results on three classical datasets demonstrate the effectiveness of EGCF in terms of recommendation accuracy and training efficiency. The code and used datasets are released at https://github.com/BlueGhostYi/ID-GRec .

Recommender systems play a crucial role in delivering personalized services to users, but the increasing volume of user data raises significant concerns about privacy, security, and utility. However, existing machine unlearning methods cannot be directly applied to recommendation systems as they overlook the collaborative information shared across users and items. More recently, a method known as RecEraser was introduced, offering partitioning and aggregation-based approaches. Nevertheless, these approaches have limitations due to their inadequate handling of additional overhead costs. In this paper, we propose A General Strategy Graph Collaborative Filtering for Recommendation Unlearning (GSGCF-RU), which is a novel model-agnostic learnable delete operator that optimizes unlearning edge consistency and feature representation consistency. Specifically, the GSGCF-RU model utilizes unlearning edge consistency to eliminate the influence of deleted elements, followed by feature representation consistency to retain knowledge after deletion. Lastly, experimental results on three real-world public benchmarks demonstrate that GSGCF-RU not only achieves efficient recommendation unlearning but also surpasses state-of-the-art methods in terms of model utility. The source code can be found at https://github.com/YongjingHao/GSGCF-RU.

Sequential Recommendation (SR) has received increasing attention due to its ability to capture user dynamic preferences. Recently, Contrastive Learning (CL) provides an effective approach for sequential recommendation by learning invariance from different views of an input. However, most existing data or model augmentation methods may destroy semantic sequential interaction characteristics and often rely on the hand-crafted property of their contrastive view-generation strategies. In this paper, we propose a Meta-optimized Seq2Seq Generator and Contrastive Learning (Meta-SGCL) for sequential recommendation, which applies the meta-optimized two-step training strategy to adaptive generate contrastive views. Specifically, Meta-SGCL first introduces a simple yet effective augmentation method called Sequence-to-Sequence (Seq2Seq) generator, which treats the Variational AutoEncoders (VAE) as the view generator and can constitute contrastive views while preserving the original sequence's semantics. Next, the model employs a meta-optimized two-step training strategy, which aims to adaptively generate contrastive views without relying on manually designed view-generation techniques. Finally, we evaluate our proposed method Meta-SGCL using three public real-world datasets. Compared with the state-of-the-art methods, our experimental results demonstrate the effectiveness of our model and the code is available.11https.//anonymous.4open.science/status/Meta-SGCL-05B5

Graph Collaborative Filtering (GCF) is designed to leverage high-order connectivity in user-item graphs, thereby significantly enhancing recommendation performance. Recent advancements have seen the integration of contrastive learning into GCF as a strategy to mitigate the challenges of data sparsity. This approach involves creating contrastive views through augmentations, followed by the generation of self-supervised signals. These signals are produced by maximizing the mutual information between the contrastive views. While this method has proven effective, we argue that current CL-based GCF models are still limited to current augmentation techniques. Existing data augmentation or noise perturbation may destroy the structural and semantic features of the original data and node attribute information is not considered. To tackle the above limitations, we propose a Meta-optimized Structure and Semantic Contrastive Learning for Graph Collaborative Filtering, named Meta-SSCL, which utilizes graph structure information and semantic information contrastive learning for recommendation. Specifically, we first model the structural and node semantic information representations with LightGCN and vanilla attention mechanism, respectively. Then consider the structural and semantic information as two contrastive views for recommendation. Next, the meta-optimized two-step training strategy generates adaptive contrastive views. Finally, we fuse structural and semantic representations for recommendation. Extensive experiments on real-world datasets demonstrate that Meta-SSCL consistently outperforms state-of-the-art sequential recommendation methods. The code is available1.

Sparse online learning has received extensive attention during the past few years. Most of existing algorithms that utilize ℓ1-norm regularization or ℓ1-ball projection assume that the feature space is fixed or changes by following explicit constraints. However, this assumption does not always hold in many real applications. Motivated by this observation, we propose a new online learning algorithm tailored for data streams described by open feature spaces, where new features can be occurred, and old features may be vanished over various time spans. Our algorithm named RSOL provides a strategy to adapt quickly to such feature dynamics by encouraging sparse model representation with an ℓ1- and ℓ2 -mixed regularizer. We leverage the proximal operator of the ℓ1,2 -mixed norm and show that our RSOL algorithm enjoys a closed-form solution at each iteration. A sub-linear regret bound of our proposed algorithm is guaranteed with a solid theoretical analysis. Empirical results benchmarked on nine streaming datasets validate the effectiveness of the proposed RSOL method over three state-of-the-art algorithms.

The user purchase behaviors are mainly influenced by their intentions (e.g., buying clothes for decoration, buying brushes for painting, etc.). Modeling a user's latent intention can significantly improve the performance of recommendations. Previous works model users' intentions by considering the predefined label in auxiliary information or introducing stochastic data augmentation to learn purposes in the latent space. However, the auxiliary information is sparse and not always available for recommender systems, and introducing stochastic data augmentation may introduce noise and thus change the intentions hidden in the sequence. Therefore, leveraging user intentions for sequential recommendation (SR) can be challenging because they are frequently varied and unobserved. In this paper, Intent contrastive learning with Cross Subsequences for sequential Recommendation (ICSRec) is proposed to model users' latent intentions. Specifically, ICSRec first segments a user's sequential behaviors into multiple subsequences by using a dynamic sliding operation and takes these subsequences into the encoder to generate the representations for the user's intentions. To tackle the problem of no explicit labels for purposes, ICSRec assumes different subsequences with the same target item may represent the same intention and proposes a coarse-grain intent contrastive learning to push these subsequences closer. Then, fine-grain intent contrastive learning is mentioned to capture the fine-grain intentions of subsequences in sequential behaviors. Extensive experiments conducted on four real-world datasets demonstrate the superior performance of the proposed ICSRec model compared with baseline methods.

Graph collaborative filtering (GCF) has achieved exciting recommendation performance with its ability to aggregate high-order graph structure information. Recently, contrastive learning (CL) has been incorporated into GCF to alleviate data sparsity and noise issues. However, most of the existing methods employ random or manual augmentation to produce contrastive views that may destroy the original topology and amplify the noisy effects. We argue that such augmentation is insufficient to produce the optimal contrastive view, leading to suboptimal recommendation results. In this article, we proposed a L earnable M odel A ugmentation C ontrastive L earning (LMACL) framework for recommendation, which effectively combines graph-level and node-level collaborative relations to enhance the expressiveness of collaborative filtering (CF) paradigm. Specifically, we first use the graph convolution network (GCN) as a backbone encoder to incorporate multi-hop neighbors into graph-level original node representations by leveraging the high-order connectivity in user-item interaction graphs. At the same time, we treat the multi-head graph attention network (GAT) as an augmentation view generator to adaptively generate high-quality node-level augmented views. Finally, joint learning endows the end-to-end training fashion. In this case, the mutual supervision and collaborative cooperation of GCN and GAT achieves learnable model augmentation. Extensive experiments on several benchmark datasets demonstrate that LMACL provides a significant improvement over the strongest baseline in terms of Recall and NDCG by 2.5%–3.8% and 1.6%–4.0%, respectively. Our model implementation code is available at https://github.com/LiuHsinx/LMACL .

Sequential recommendation aims to predict the next item of interest to users based on their historical behavior data. Usually, users’ global and local preferences jointly affect the final recommendation result in different ways. Most existing works use transformers to globally model sequences, which makes them face the dilemma of quadratic computational complexity when dealing with long sequences. Moreover, the scope setting of the user’s local preference is usually static and single, and cannot cover richer multi-level local semantics. To this end, we proposed a parallel architecture for capturing global representation and M ulti-granularity L ocal dependencies with M LP for sequential Rec ommendation ( MLM4Rec ). For global representation, we utilize modified MLP-Mixer to capture global information of user sequences due to its simplicity and efficiency. For local representation, we incorporate convolution into MLP and propose a multi-granularity local awareness mechanism for capturing richer local semantic information. Moreover, we introduced a weight pooling method to adaptively fuse local-global representations instead of directly concatenation. Our model has the advantages of low complexity and high efficiency thanks to its simple MLP structure. Experimental results on three public datasets demonstrate the effectiveness of our proposed model. Our code is available here 1 .

Local spatial-aware community detection refers to detecting a spatial-aware community for a given node using local information. A spatial-aware community means that nodes in the community are tightly connected in structure, and their locations are close to each other. Existing studies focus on detecting the local non-overlapping spatial-aware community, i.e., detecting a spatial-aware community containing the given node. However, many geosocial networks often contain overlapping spatial-aware communities. Therefore, we propose a local overlapping spatial-aware community detection (LOSCD) problem, which aims to detect all spatial-aware communities that contain a given node with local information. To address LOSCD problem, we design an algorithm based on Spatial Modularity and Edge Similarity, called SMES. SMES contains two processes: spatial expansion and structure detection. The spatial expansion process involves using spatial modularity to identify nodes that are spatially close, while the structural detection process employs edge similarity to identify nodes that are structurally close. Experimental results demonstrate that SMES outperforms comparison algorithms in terms of both structural and spatial cohesiveness.

Object detection is a widely studied problem in existing works. However, in this paper, we turn to a more challenging problem of “ Covered Object Reasoning ”, aimed at reasoning the category label of target object in the given image particularly when it has been totally covered (or invisible ). To resolve this problem, we propose CoBjeason to seize the opportunity when visual reasoning meets the knowledge graph, where “ empirical cognition ” on common visual contexts have been incorporated as knowledge graph to conduct reinforced multi-hop reasoning via two collaborative agents. Such two agents, for one thing, stand at the covered object (or unknown entity ) to observe the surrounding visual cues in the given image and gradually select entities and relations from the global gallery-level knowledge graph which contains entity-pairs frequently occurring across the entire image-collection, so as to infer the main structure of image-level knowledge graph forward expanded from the unknown entity . In turn, for another, based on the reasoned image-level knowledge graph, the semantic context among entities will be aggregated backward into unknown entity to select an appropriate entity from the global gallery-level knowledge graph as the reasoning result. Moreover, such two agents will collaborate with each other, securing that the above Forward & Backward Reasoning will step towards the same destination of the higher performance on covered object reasoning. To our best knowledge, this is the first work on Covered Object Reasoning with Knowledge Graphs and reinforced Multi-Agent collaboration. Particularly, our study on Covered Object Reasoning and the proposed model CoBjeason could offer novel insights into more basic Computer Vision (CV) tasks, such as Semantic Segmentation with better understanding on the current scene when some objects are blurred or covered, Visual Question Answering with enhancement on the inference in more complicated visual context when some objects are covered or invisible, and Image Caption Generation with the augmentation on the richness of visual context for images containing partially visible objects. The improvement on the above basic CV tasks can further refine more complicated ones involved with nuanced visual interpretation like Autonomous Driving, where the recognition and reasoning on partially visible or covered object are critical. According to the experimental results, our proposed CoBjeason can achieve the best overall ranking performance on covered object reasoning compared with other models, meanwhile enjoying the advantage of lower “ exploration cost ”, with the insensitivity against the long-tail covered objects and the acceptable time complexity.

Online shopping has become a crucial way to encourage daily consumption, where the User-generated, or crowdsourced product comments, can offer a broad range of feedback on e-commerce products. As a result, integrating critical opinions or major attitudes from the crowdsourced comments can provide valuable feedback for marketing strategy adjustment or product-quality monitoring. Unfortunately, the scarcity of annotated ground truth on the integrated comment, or the limited gold integration reference, has incurred the infeasibility of the regular supervised-learning-based comment integration. To resolve this problem, in this article, inspired by the principle of Transfer Learning, we propose a three-stage transferable and generative crowdsourced comment integration framework ( TTGCIF ) based on zero-and-few-shot learning with the support of domain distribution alignment. The proposed framework aims at generating abstractive integrated comment in target domain via the enhanced neural text generation model, by referring the available integration resource in related source domains, to avoid the exhausted effort on resource annotation devoted to the target domain. Specifically, at the first stage, to enhance the domain transferability, representations on the crowdsourced comments have been aligned up between the source and target domain, by minimizing the domain distribution discrepancy in the kernel space. At the second stage, Zero-shot comment integration mechanism has been adopted to deal with the dilemma that none of the gold integration reference may be available in target domain. In other words, taking the sample-level semantic prototype as input, the enhanced neural text generation model in TTGCIF is trained to learn data semantic association among different domains via semantic prototype transduction, so that the “ unlabeled ” crowdsourced comments in target domain can be associated with existing integration references in related source domains. At the third stage, based on the parameters trained at the second stage, fast domain adaptation mechanism in a Few-shot manner has also been adopted by seeking most potential parameters along the gradient direction constrained by instances across multiple source domains. In this way, parameters in TTGCIF can be sensitive to any alteration on training data, ensuring that even if only few annotated resource in target domain are available for “Fine-tune,” TTGCIF can still react promptly to achieve effective target domain adaptation. According to the experimental results, TTGCIF can achieve the best transferable product comment integration performance in target domain, with fast and stable domain adaption effect depending on no more than 10% annotated resource in target domain. More importantly, even if TTGCIF has not been fine-tuned on the target domain, yet by referring to the available integration resource in related source domains, the integrated comments generated by TTGCIF on the target domain are still superior to those generated by models already fine-tuned on the target domain.

Contrastive learning with Transformer-based sequence encoder has gained predominance for sequential recommendation due to its ability to mitigate the data noise and the data sparsity issue. However, existing contrastive learning approaches for sequential recommendation still suffer from two limitations. First, they mainly center on left-to-right unidirectional Transformers as base encoders, which are suboptimal for sequential recommendation because user behaviors may not be a rigid left-to-right sequence. Second, they devise contrastive learning objectives only from the sequence level, neglecting the rich self-supervision signals from the feature level. To address these limitations, we propose a novel framework called Feature-aware Contrastive Learning with bidirectional Transformers for sequential Recommendation (FCLRec) to effectively leverage feature information for sequential recommendation. Specifically, we first augment bidirectional Transformers with a novel feature-aware self-attention module that is able to simultaneously model the complex relationships between sequences and features. Next, we propose a novel feature-aware contrastive learning objective that generates a collection of positive samples via three types of augmentations from three different levels. Finally, we adopt feature prediction as an auxiliary task to strengthen the connections between items and features. Our experimental results on four public benchmark datasets show that FCLRec outperforms the state-of-the-art methods for sequential recommendation.

Sequential Recommendation (SR) methods play a crucial role in recommender systems, which aims to capture users' dynamic interest from their historical interactions. Recently, Contrastive Learning (CL), which has emerged as a successful method for sequential recommendation, utilizes various data augmentations to generate contrastive views to mine supervised signals from data to alleviate data sparsity issues. However, most existing sequential data augmentation methods may destroy semantic sequential interaction characteristics. Meanwhile, they often adopt random operations when generating contrastive views leading to suboptimal performance. To this end, in this paper, we propose a Learnable Model Augmentation Contrastive learning for sequential Recommendation (LMA4Rec). Specifically, LMA4Rec first takes the model-based augmentation method to generate constructive views. Then, LMA4Rec uses Learnable Bernoulli Dropout (LBD) to implement learnable model augmentation operations. Next, contrastive learning is used between the contrastive views to extract supervised signals. Furthermore, a novel multi-positive contrastive learning loss alleviates the supervised sparsity issue. Finally, experiments on public datasets show that our LMA4Rec method effectively improved sequential recommendation performance compared with the state-of-the-art baseline methods.

Owing to the lack of a universal definition of communities, some semi-supervised community detection approaches learn the concept of community structures from known communities, and then dig out communities using learned concepts of communities. In some cases, users are only interested in the community containing a given node. However, communities detected by these semi-supervised approaches may not contain a given node. Besides, these methods traverse the entire network to detect many communities and cost more resources than a local algorithm. Therefore, it is necessary and meaningful to find the local community that contains a given node with prior information on the local network around the given node. We call this a Semi-supervised Local Community Detection (SLCD) problem. In this paper, prior information refers to certain known communities. To address the SLCD problem, we propose the Semi-supervised Local community detection with the Structural Similarity algorithm, called SLSS, which uses some known communities instead of all known communities. The idea of SLSS is to use the structural similarity between the known communities and the detected community, calculated by the graph kernel, to guide the expansion of the community. Experimental results show that SLSS outperforms other algorithms on six real-world datasets.

The identification of the local community structure in geo-social networks has been gaining increasing attention. The structure of geo-social networks evolves over time with the addition/deletion of edges/nodes and the update of node locations, which has motivated recent studies to mine local communities in dynamic geo-social networks. Mining communities in evolving geo-social networks is essential for understanding the evolution of group behaviors. However, in most previous studies on the community mining in dynamic networks, local spatial-aware communities were not identified in evolving geo-social networks. Therefore, in this study, the problem of determining local spatial-aware communities in evolving geo-social networks is proposed. To address this problem, we propose a parameter-free algorithm, called LSADEN. Specifically, LSADEN involves two main steps: i) selecting candidate nodes, where LSADEN defines the community dominance relation under dynamic environments to obtain candidate nodes that improve the community in terms of the community quality or the smoothness between communities at adjacent time stamps; ii) community expansion, where LSADEN designs the Manhattan distance of communities to add some candidate nodes to the local community. Experimental results on six real-world datasets and one synthetic dataset show that LSADEN performs well both in terms of the quality of communities and the smoothness between communities at adjacent time stamps.

Nowadays, crowdsourced documents like Wikipedia pages and comments on products are all over the Internet. However, documents generated by crowdsourcing participants may contain inconsistent facts, implicit semantics and fabricated contents, thus threatening the trustworthiness of information content security available in Internet. To address this problem, we propose FuFaction, enabled by an enhanced observation mechanism based on the notion of hybrid-mask consisting of a hard-mask and a soft-mask, to eliminate factual inconsistencies on crowdsourced documents at the hidden-state level (or in a fuzzy way), according to the given evidence retrieved from an external open domain. Specifically, instead of focusing on a specific category of factual inconsistency, FuFaction captures anomalous hidden-states between a crowdsourced document and evidence obtained via a reverse-attention mechanism, where a hard-mask controls the attending direction as bidirectional and unidirectional for better understanding on semantics. Then, a soft-mask is generated with the help of the hard-masked reverse-attention to revise or mask anomalous hidden-states on the crowdsourced document. Afterwards, the masked hidden-states are further refined by a cross reverse-attention and factual consistency reinforcement strategy, based on which a new crowdsourced document with higher factual consistency is generated via neural text generation. According to our experimental results, FuFaction can effectively deal with the fuzzy factual inconsistencies on crowdsourced documents, achieving the overall best performance in terms of factual consistency metrics with a little higher (yet still competitive) editing cost on literal vocabulary, so as to reflect factually consistent semantics supported by the given evidence.

Deep learning has been widely adopted in Android malicious software (malware) detection. However, poor explanation in deep learning-based detection models severely undermines user trusts and poses a significant obstacle to their practical promotion in critical security domains. Some studies strive to uncover the rationale behind a model's decision. Unfortunately, these efforts are often hindered by the limitations of feature extraction methods, such as primarily relying on static analysis to derive separate and approximate behavioral descriptions of applications (apps). As a result, establishing a reliable interpretation for deep learning-based malware detection models remains an open issue. In this work, we propose a novel framework XDeepMal to interpret deep learning-based malware detection models. Specifically, in XDeepMal, we formulate a dynamic analysis tool XTracer+to capture runtime behaviors of apps and automatically generate their continuous behavior trajectories. Then, we propose a novel interpreter to pinpoint certainty behavior fragments that are crucial for deep learning models to make their decisions. This approach regards the identification of the most critical fragments as an optimization problem and leverages heuristic algorithms for implementation. We conduct extensive experiments on a real-world dataset to investigate the effectiveness and reliability of XDeepMal. These experiments cover intuitive case studies (malware family and individual app) and in-depth quantitative analysis. Additionally, we evaluate its coverage and efficiency. Our experimental results demonstrate that XDeepMal is capable of generating convincing interpretations for deep learning (e.g., Transformer) based models within feasible inference time, which greatly benefits security analysts in accurately comprehending why an app is identified as malware by deep learning-based detection models.

Online learning updates models incrementally with new data, avoiding large storage requirements and costly model recalculations. In this paper, we introduce “OLR-WA; OnLine Regression with Weighted Average”, a novel and versatile multivariate online linear regression model. We also investigate scenarios involving drift, where the underlying patterns in the data evolve over time, conduct convergence analysis, and compare our approach with existing online regression models. The results of OLR-WA demonstrate its ability to achieve performance comparable to the batch regression, while also showcasing comparable or superior performance when compared with other state-of-the-art online models, thus establishing its effectiveness. Moreover, OLR-WA exhibits exceptional performance in terms of rapid convergence, surpassing other online models with consistently achieving high r2values as a performance measure from the first iteration to the last iteration, even when initialized with minimal amount of data points, as little as 1% to 10% of the total data points. In addition to its ability to handle time-based (temporal drift) scenarios, remarkably, OLR-WA stands out as the only model capable of effectively managing confidence-based challenging scenarios. It achieves this by adopting a conservative approach in its updates, giving priority to older data points with higher confidence levels. In summary, OLR-WA’s performance further solidifies its versatility and utility across different contexts, making it a valuable solution for online linear regression tasks.

The rapid advancement of neural networks has permeated industries worldwide. However, this remarkable success often faces adversity while deploying on devices with limited resources due to the high computational power demands and storage requirements. To address these challenges, a model compressing technique such as knowledge distillation (KD) has emerged as an effective process for training compact models. However, due to the lack of interpretability, knowledge distillation is often hard to explain. In this paper, our focus is to implement a knowledge distillation framework that can provide interpretability alongside improved performance. With the help of our experiments, we demonstrate that this can be achieved by integrating first-order logical formulas into a neurosymbolic learning approach within the knowledge distillation framework. Diverging from the prior research, in this paper, we introduce a Neurosymbolic Knowledge Distillation Framework (KD-LTN), which is a composition of Logic Tensor Network (LTN) and Knowledge Distillation (KD). Notably, our KD-LTN network not only enhances interpretability but also achieves accuracy improvements, compared with the conventional knowledge distillation framework.

Simplicial neural networks (SNNs) were proposed to generate higher-order simplicial complex representations as vectors that encode not only pairwise relationships but also higher-order interactions between nodes. Although these vectors allowing us to consider richer data representations compared to typical graph convolution, most real-world graphs associated with molecule or human-related activities are often sensitive and might contain confidential information, e.g., molecular geometry or friend lists. However, little works investigate the potential threats for these simplicial complexes (higher-order interactions between nodes). We name this threat by Simplicial Complexes Reconstruction Attack (SCRA) and conduct this attack by studying whether the vectors can be inverted to (approximately) recover the simplicial complexes who used to generate them. Specifically, we first generate the vectors via a k-simplex2vec approach that extends the node2vec algorithm to simplices of higher dimensions to associate Euclidean vectors to simplicial complexes. We then present a Simplex2vec Backward algorithm to perform the SCRA on k-simplex2vec vectors by pointwise mutual information (PMI) matrix reconstruction.

Releasing the representations of nodes in real-world graphs associated with people or human-related activities, such as social and economic networks, gives adversaries a potential way to infer the sensitive information of edges. For example, graph convolutional layers initially aggregate node representations with their neighbors before passing them through non-linear activation functions. Hence, the released node representations may potentially breach edge privacy of the node neighbors. Thus, in this work, we study whether representations can be inverted to recover the graph used to generate them. We study three types of outputs that are trained on the graph, i.e., representations output from graph convolutional networks (GCNs), representations output from graph attention networks (GATs), and representations output from our proposed simplicial neural networks (SNNs). Unlike the first two types of representations that only encode pairwise relationships, the third type of representation, i.e., SNN outputs, encodes higher-order interactions (e.g., homological features) between nodes. We propose two graph reconstruction attacks (GRAs), i.e., Type-1 and Type-2 attacks, to recover a graph’s adjacency matrix from the three types of outputs trained on the graph. Specifically, our GRAs utilize a graph-decoder to minimize the reconstruction loss for the generated adjacency matrix via back-propagation. Our conclusions are two folds. First, our Type-2 attack achieves the best performance among all current GRAs. Second, we find that GCN outputs obtain the least precision and AUC on five datasets, followed by the GAT outputs, followed by the SNN outputs. Therefore, the SNN outputs reveal the lowest privacy-preserving ability to defend the GRAs. We further propose an unbiased multi-bit rectifier, by which the server can communicate with the nodes to privately collect their representations to defend the GRAs from potential adversaries.

Sequential recommendation (SR) aims to model user preferences by capturing behavior patterns from their item historical interaction data. Most existing methods model user preference in the time domain, omitting the fact that users’ behaviors are also influenced by various frequency patterns that are difficult to separate in the entangled chronological items. However, few attempts have been made to train SR in the frequency domain, and it is still unclear how to use the frequency components to learn an appropriate representation for the user. To solve this problem, we shift the viewpoint to the frequency domain and propose a novel Contrastive Enhanced SLIde Filter MixEr for Sequential Recommendation, named SLIME4Rec. Specifically, we design a frequency ramp structure to allow the learnable filter slide on the frequency spectrums across different layers to capture different frequency patterns. Moreover, a Dynamic Frequency Selection (DFS) and a Static Frequency Split (SFS) module are proposed to replace the self-attention module for effectively extracting frequency information in two ways. DFS is used to select helpful frequency components dynamically, and SFS is combined with the dynamic frequency selection module to provide a more fine-grained frequency division. Finally, contrastive learning is utilized to improve the quality of user embedding learned from the frequency domain. Extensive experiments conducted on five widely used benchmark datasets demonstrate our proposed model performs significantly better than the state-of-the-art approaches. Our code is available at https://github.com/sudaada/SLIME4Rec.

Crowdsourcing is a sourcing model where individuals or organizations obtain goods and services from a large, relatively open and often rapidly evolving group of internet users. The most common way that crowdsourcing can facilitate machine learning is to annotate instances with labels [1] . However, the same instance may have inconsistent class labels, in the eyes of various annotators. Therefore, current efforts in crowdsourcing mainly focus on the truth inference or label integration, to remove inconsistent labels or to alleviate biased labeling. In turn, instances with the integrated labels could facilitate the training on machine learning models. The future direction of crowdsourcing is to apply more fine-grained truth inference methods to different application domains [2] . Consequently, we evolve toward another challenging problem of comment integration. That is, how can we integrate or summarize the core opinions of multiple product comments obtained from users, rather than the discrete labels.

The self-attention mechanism, which equips with a strong capability of modeling long-range dependencies, is one of the extensively used techniques in the sequential recommendation field. However, many recent studies represent that current self-attention based models are low-pass filters and are inadequate to capture high-frequency information. Furthermore, since the items in the user behaviors are intertwined with each other, these models are incomplete to distinguish the inherent periodicity obscured in the time domain. In this work, we shift the perspective to the frequency domain, and propose a novel Frequency Enhanced Hybrid Attention Network for Sequential Recommendation, namely FEARec. In this model, we firstly improve the original time domain self-attention in the frequency domain with a ramp structure to make both low-frequency and high-frequency information could be explicitly learned in our approach. Moreover, we additionally design a similar attention mechanism via auto-correlation in the frequency domain to capture the periodic characteristics and fuse the time and frequency level attention in a union model. Finally, both contrastive learning and frequency regularization are utilized to ensure that multiple views are aligned in both the time domain and frequency domain. Extensive experiments conducted on four widely used benchmark datasets demonstrate that the proposed model performs significantly better than the state-of-the-art approaches.

Sequential recommendation aims to capture users' dynamic interest and predicts the next item of users' preference. Most sequential recommendation methods use a deep neural network as sequence encoder to generate user and item representations. Existing works mainly center upon designing a stronger sequence encoder. However, few attempts have been made with training an ensemble of networks as sequence encoders, which is more powerful than a single network because an ensemble of parallel networks can yield diverse prediction results and hence better accuracy. In this paper, we present Ensemble Modeling with contrastive Knowledge Distillation for sequential recommendation (EMKD). Our framework adopts multiple parallel networks as an ensemble of sequence encoders and recommends items based on the output distributions of all these networks. To facilitate knowledge transfer between parallel networks, we propose a novel contrastive knowledge distillation approach, which performs knowledge transfer from the representation level via Intra-network Contrastive Learning (ICL) and Cross-network Contrastive Learning (CCL), as well as Knowledge Distillation (KD) from the logits level via minimizing the Kullback-Leibler divergence between the output distributions of the teacher network and the student network. To leverage contextual information, we train the primary masked item prediction task alongside the auxiliary attribute prediction task as a multi-task learning scheme. Extensive experiments on public benchmark datasets show that EMKD achieves a significant improvement compared with the state-of-the-art methods. Besides, we demonstrate that our ensemble method is a generalized approach that can also improve the performance of other sequential recommenders. Our code is available at this link: https://github.com/hw-du/EMKD.

Contrastive Learning (CL) performances as a rising approach to address the challenge of sparse and noisy recommendation data. Although having achieved promising results, most existing CL methods only perform either hand-crafted data or model augmentation for generating contrastive pairs to find a proper augmentation operation for different datasets, which makes the model hard to generalize. Additionally, since insufficient input data may lead the encoder to learn collapsed embeddings, these CL methods expect a relatively large number of training data (e.g., large batch size or memory bank) to contrast. However, not all contrastive pairs are always informative and discriminative enough for the training processing. Therefore, a more general CL-based recommendation model called Meta-optimized Contrastive Learning for sequential Recommendation (MCLRec) is proposed in this work. By applying both data augmentation and learnable model augmentation operations, this work innovates the standard CL framework by contrasting data and model augmented views for adaptively capturing the informative features hidden in stochastic data augmentation. Moreover, MCLRec utilizes a meta-learning manner to guide the updating of the model augmenters, which helps to improve the quality of contrastive pairs without enlarging the amount of input data. Finally, a contrastive regularization term is considered to encourage the augmentation model to generate more informative augmented views and avoid too similar contrastive pairs within the meta updating. The experimental results on commonly used datasets validate the effectiveness of MCLRec.

Sequential recommendation has been a widely popular topic of recommender systems. Existing works have contributed to enhancing the prediction ability of sequential recommendation systems based on various methods, such as recurrent networks and self-attention mechanisms. However, they fail to discover and distinguish various relationships between items, which could be underlying factors which motivate user behaviors. In this article, we propose an Edge-Enhanced Global Disentangled Graph Neural Network (EGD-GNN) model to capture the relation information between items for global item representation and local user intention learning. At the global level, we build a global-link graph over all sequences to model item relationships. Then a channel-aware disentangled learning layer is designed to decompose edge information into different channels, which can be aggregated to represent the target item from its neighbors. At the local level, we apply a variational auto-encoder framework to learn user intention over the current sequence. We evaluate our proposed method on three real-world datasets. Experimental results show that our model can get a crucial improvement over state-of-the-art baselines and is able to distinguish item features.

Exploiting social networks is expected to enhance the performance of recommender systems when interaction information is sparse. Existing social recommendation models focus on modeling multi-graph structures and then aggregating the information from these multiple graphs to learn potential user preferences. However, these methods often employ complex models and redundant parameters to get a slight performance improvement. Contrastive learning has been widely researched as an effective paradigm in the area of recommendation. Most existing contrastive learning-based models usually focus on constructing multi-graph structures to perform graph augmentation for contrastive learning. However, the effect of graph augmentation on contrastive learning is inconclusive. In view of these challenges, in this work, we propose a contrastive learning based graph convolution network for social recommendation (CLSR), which integrates information from both the social graph and the interaction graph. First, we propose a fusion-simplified method to combine the social graph and the interaction graph. Technically, on the basis of exploring users’ interests by interaction graph, we further exploit social connections to alleviate data sparsity. By combining the user embeddings learned through two graphs in a certain proportion, we can obtain user representation at a finer granularity. Meanwhile, we introduce a contrastive learning framework for multi-graph network modeling, where we explore the feasibility of constructing positive and negative samples of contrastive learning by conducting data augmentation on embedding representations. Extensive experiments verify the superiority of CLSR’s contrastive learning framework and fusion-simplified method of integrating social relations.

Sequential recommendation, which aims to recommend next item that the user will likely interact in a near future, has become essential in various Internet applications. Existing methods usually consider the transition patterns between items, but ignore the transition patterns between features of items. We argue that only the item-level sequences cannot reveal the full sequential patterns, while explicit and implicit feature-level sequences can help extract the full sequential patterns. Meanwhile, the item-level sequential recommendation also suffers from limited supervised signal issues. In this article, we propose a novel model Feature-level Deeper Self-Attention Network with Contrastive Learning (FDSA-CL) for sequential recommendation. Specifically, FDSA-CL first integrates various heterogeneous features of items into feature-level sequences with different weights through a vanilla attention mechanism. After that, FDSA-CL applies separated self-attention blocks on item-level sequences and feature-level sequences, respectively, to model item transition patterns and feature transition patterns. Moreover, we propose contrastive learning and item feature recommendation tasks to capture the embedding commonality and further utilize the beneficial interaction among the two levels, so as to alleviate the sparsity of the supervised signal and extract the most critical information. Finally, we jointly optimize the above tasks. We evaluate the proposed model using two real-world datasets and experimental results show that our model significantly outperforms the state-of-the-art approaches.

With the explosive growth of short texts on users' preferences, learning discriminative and coherent latent topics from short texts is a critical work, since many practical applications require semantic understandings that short texts convey explicitly and implicitly. However, existing short text topic learning methods face the challenge of fully capturing semantically related co-occurrence phrases. Therefore, this paper proposes a novel Heterogeneous Information Network-based Short Text Topic learning approach (HIN-ShoTT) in terms of parts of speech, without depending on any auxiliary information. Specifically, HIN-ShoTT can be decomposed into three phases: i) seeking semantic relations among words, where HIN-ShoTT models multiple semantic relations among words based on a Heterogeneous Information Network (HIN) in terms of parts of speech; ii) extracting co-occurrence phrases and filtering noises, where HIN-ShoTT defines parts-of-speech meta structures to guide co-occurrence phrase extraction and a self-adapting threshold filtering module is proposed for discarding noises; and iii) inferring topics, where HIN-ShoTT models the generative process of co-occurrence phrases to make topic learning effective with the abundant corpus-level information. Our experimental results on three real-world datasets not only show that HIN-ShoTT performs well, but also demonstrate that it is feasible to incorporate HIN into short text topic learning for accuracy improvement.

Effective streaming feature selection in dynamic on-line environments is essential in numerous applications. However, most existing methods evaluate high-dimensional features individually and ignore the potentially pertainable group structures of features. Moreover, the class imbalance underlying streaming data may further decrease the discriminative efficacy of the selected features, resulting in deteriorated classification performance. Motivated by this observation, we propose a proximal cost-sensitive sparse group online learning (PCSGOL) framework to handle imbalanced and high-dimensional streaming data. Specifically, we formulate this issue as a new cost-sensitive online optimization problem by leveraging the ℓ2-norm, ℓ1-norm, and group-wise sparsity constraints in the dual averaging regularization. The average weighted distance is also introduced in PCSGOL to achieve stable prediction results. We mathematically derive closed-form solutions to the optimization problems with four modified hinge loss functions, leading to four variants of PCSGOL. Extensive empirical studies on real-world streaming datasets demonstrate the effectiveness of our proposed method.

Contrastive learning with Transformer-based sequence encoder has gained predominance for sequential recommendation. It maximizes the agreements between paired sequence augmentations that share similar semantics. However, existing contrastive learning approaches in sequential recommendation mainly center upon left-to-right unidirectional Transformers as base encoders, which are suboptimal for sequential recommendation because user behaviors may not be a rigid left-to-right sequence. To tackle that, we propose a novel framework named Contrastive learning with Bidirectional Transformers for sequential recommendation (CBiT). Specifically, we first apply the slide window technique for long user sequences in bidirectional Transformers, which allows for a more fine-grained division of user sequences. Then we combine the cloze task mask and the dropout mask to generate high-quality positive samples and perform multi-pair contrastive learning, which demonstrates better performance and adaptability compared with the normal one-pair contrastive learning. Moreover, we introduce a novel dynamic loss reweighting strategy to balance between the cloze task loss and the contrastive loss. Experiment results on three public benchmark datasets show that our model outperforms state-of-the-art models for sequential recommendation. Our code is available at this link: https://github.com/hw-du/CBiT/tree/master.

Most existing online learning methods focus on mining ever-evolving streaming data based on the principle of first-order optimization. However, one drawback of these methods is the slow convergence rate in each iteration, resulting in sub-optimal solutions and deteriorated performance. Second-order methods, while are able to provide faster convergence, have been under-studied due to the high cost of computing the curvature information. To address this problem, in this paper, we develop a second-order projection dual averaging based online learning (SPDA) method to effectively handle high-throughput streaming data. By fully exploiting the regularized dual averaging optimization, the second-order information, and an optimal projection operator, SPDA converges fast with fairly optimal solutions. Two speed-up versions of SPDA, i.e., SPDA-diag and SPDA-sketch, are developed via the diagonal operator and Hessian sketch, respectively. Theoretical derivations on the regret bound of SPDA establish a solid convergence guarantee for this method. Extensive experiments demonstrate the efficacy of the proposed algorithms on large-scale online learning tasks, such as online binary and multi-class classification and online anomaly detection, shedding light on their potential wide applications.

Crowdsourcing is a hotspot research field which can facilitate machine learning by collecting labels to train models. Consequently, the state-of-the-art research efforts in crowdsourcing focus on truth inference or label integration, to remove inconsistent labels or to alleviate biased labeling. In turn, the integrated labels will be used to fine-tune machine learning models. Particularly, in this paper, we change the target of truth inference in crowdsourcing from discrete labels to multiple comments given by online participants, that is, the integration of the crowdsourced comments. For such a goal, we propose aSelf-play andSentiment-EmphasizedCommentIntegrationFramework (SSECIF), based on deepQ-learning, with three unique features. First, our framework SSECIF can generate the comment integration in a totally self-play way, without relying on the ground truth generated by human effort. Second, the integrated comment generated by SSECIF can include salient content with low redundancy. Third, the proposed framework SSECIF has emphasized, with a higher intensity, the sentiment in the integrated comment, in order to reflect the attitude or opinion more obviously. Extensive evaluation on real-world datasets demonstrates that SSECIF has achieved the best overall performance in terms of both effectiveness and efficiency, compared with the state-of-the-art methods.

Next Point-of-Interest (POI) recommendation which is of great value to both users and POI holders is a challenging task since complex sequential patterns and rich contexts are contained in extremely sparse user check-in data. Recently proposed embedding techniques have shown promising results in alleviating the data sparsity issue by modeling context information, and Recurrent Neural Network (RNN) has been proved effective in the sequential prediction. However, existing next POI recommendation approaches train the embedding and network model separately, which cannot fully leverage rich contexts. In this paper, we propose a novel unified neural network framework, named NeuNext, which leverages POI context prediction to assist next POI recommendation by joint learning. Specifically, the Spatio-Temporal Gated Network (STGN) is proposed to model personalized sequential patterns for users’ long and short term preferences in the next POI recommendation. In the POI context prediction, rich contexts on POI sides are used to construct graph, and enforce the smoothness among neighboring POIs. Finally, we jointly train the POI context prediction and the next POI recommendation to fully leverage labeled and unlabeled data. Extensive experiments on real-world datasets show that our method outperforms other approaches for next POI recommendation in terms of Accuracy and MAP.

Android is a growing target for malicious software (malware) because of its popularity and functionality. Malware poses a serious threat to users’ privacy, money, equipment and file integrity. A series of data-driven malware detection methods were proposed. However, there exist two key challenges for these methods: (1) how to learn effective feature representation from raw data; (2) how to reduce the dependence on the prior knowledge or human labors in feature learning. Inspired by the success of deep learning methods in the feature representation learning community, we propose a malware detection framework which starts with learning rich-features by a novel unsupervised feature learning algorithm Merged Sparse Auto-Encoder (MSAE). In order to extract more compact and discriminative feature from the rich-features to further boost the malware detection capability, a hybrid deep network learning algorithm Stacked Hybrid Learning MSAE and SDAE (SHLMD) is established by further incorporating a classical deep learning method Stacked Denoising Auto-encoders (SDAE). After that, we feed the feature learned by MSAE and SHLMD respectively to classification algorithms, e.g., Support Vector Machine (SVM) or K-NearestNeighbor (KNN), to train a malware detection model. Evaluation results on two real-world datasets demonstrate that SHLMD achieves 94.46 and 90.57 percent accuracy respectively, which outperforms the classical unsupervised feature representation learning Sparse Auto-encoder (SAE). MSAE performs similarly to SAE. SHLMD can further improve the performance of MSAE and the supervised fine-tuned method SDAE. Besides, we compare the performance of our methods with that of state-of-the-art detection approaches, including classical deep-learning-based methods. Extensive experiments show that our proposed methods are effective enough to detect Android malware.

For decades, research in natural language processing (NLP) has focused on summarization. Sequence-to-sequence models for abstractive summarization have been studied extensively, yet generated summaries commonly suffer from fabricated content, and are often found to be near-extractive. We argue that, to address these issues, summarizers need to acquire the co-references that form multiple types of relations over input sentences, e.g., 1-to-N, N-to-1, and N-to-N relations, since the structured knowledge for text usually appears on these relations. By allowing the decoder to pay different attention to the input sentences for the same entity at different generation states, the structured graph representations generate more informative summaries. In this paper, we propose a hierarchical graph attention networks (HGATs) for abstractive summarization with a topic-sensitive PageRank augmented graph. Specifically, we utilize dual decoders, a sequential sentence decoder, and a graph-structured decoder (which are built hierarchically) to maintain the global context and local characteristics of entities, complementing each other. We further design a greedy heuristic to extract salient users' comments while avoiding redundancy to drive a model to better capture entity interactions. Our experimental results show that our models produce significantly higher ROUGE scores than variants without graph-based attention on both SSECIF and CNN/Daily Mail (CNN/DM) datasets.

Private information can either take the form of key phrases that are explicitly contained in the text or be implicit. For example, demographic information about the author of a text can be predicted with above-chance accuracy from linguistic cues in the text itself. Letting alone its explicitness, some of the private information correlates with the output labels and therefore can be learned by a neural network. In such a case, there is a tradeoff between the utility of the representation (measured by the accuracy of the classification network) and its privacy. This problem is inherently a multi-objective problem because these two objectives may conflict, necessitating a trade-off. Thus, we explicitly cast this problem as multi-objective optimization (MOO) with the overall objective of finding a Pareto stationary solution. We, therefore, propose a multiple-gradient descent algorithm (MGDA) that enables the efficient application of the Frank-Wolfe algorithm [10] using the line search. Experimental results on sentiment analysis and part-of-speech (POS) tagging show that MGDA produces higher-performing models than most recent proxy objective approaches, and performs as well as single objective baselines.

Sequential recommendation has become an attractive topic in recommender systems. Existing sequential recommendation methods, including the methods based on the state-of-the-art self-attention mechanism, usually employ deterministic neural networks to represent user preferences as fixed-points in the latent feature spaces. However, the fixed-point vector lacks the ability to capture the uncertainty and dynamics of user preferences that are prevalent in recommender systems. In this paper, we propose a new Variational Self-Attention Network (VSAN), which introduces a variational autoencoder (VAE) into the self-attention network to capture latent user preferences. Specifically, we represent the obtained self-attention vector as density via variational inference, whose variance well characterizes the uncertainty of user preferences. Furthermore, we employ self-attention networks to learn the inference process and generative process of VAE, which well captures long-range and local dependencies. Finally, we evaluate our proposed method VSAN with two public real-world datasets. Our experimental results show the effectiveness of our model compared to the state-of-the-art approaches.

Recent studies have shown that graph learning models are highly vulnerable to adversarial attacks, and network alignment methods are no exception. How to enhance the robustness of network alignment against adversarial attacks remains an open research problem. In this paper, we propose a robust network alignment solution, RNA, for offering preemptive protection of existing network alignment algorithms, enhanced with the guidance of effective adversarial attacks. First, we analyze how popular iterative gradient-based adversarial attack techniques suffer from gradient vanishing issues and show a fake sense of attack effectiveness. Based on dynamical isometry theory, an attack signal scaling (ASS) method with established upper bound of feasible signal scaling is introduced to alleviate the gradient vanishing issues for effective adversarial attacks while maintaining the decision boundary of network alignment. Second, we develop an adversarial perturbation elimination (APE) model to neutralize adversarial nodes in vulnerable space to adversarial-free nodes in safe area, by integrating Dirac delta approximation (DDA) techniques and the LSTM models. Our proposed APE method is able to provide proactive protection to existing network alignment algorithms against adversarial attacks. The theoretical analysis demonstrates the existence of an optimal distribution for the APE model to reach a lower bound. Last but not least, extensive evaluation on real datasets presents that RNA is able to offer the preemptive protection to trained network alignment methods against three popular adversarial attack models.

Recently accumulated massive amounts of geo-tagged photos provide an excellent opportunity to understand human behaviors and can be used for personalized tour recommendation. However, no existing work has considered the visual content information in these photos for tour recommendation. We believe the visual features of photos provide valuable information on measuring user / Point-of-Interest (POI) similarities, which is challenging due to data sparsity. To this end, in this paper, we propose a visual feature enhanced tour recommender system, named ‘Photo2Trip’, to utilize the visual contents and collaborative filtering models for recommendation. Specifically, we propose a Visual-enhanced Probabilistic Matrix Factorization model (VPMF), which integrates visual features into the collaborative filtering model, to learn user interests by leveraging the historical travel records. We then extend VPMF to End-to-End training framework to incorporate users (POIs) latent factors into the learning process of the visual content of photos, which generalizes the applicability of the proposed VPMF framework in tour recommendation. Extensive empirical studies verify that our proposed visual-enhanced personalized tour recommendation method outperforms other benchmark methods in terms of recommendation accuracy. The results also show that visual features are effective in alleviating the data sparsity and cold start problems on personalized tour recommendation.

Chronic alcohol consumption has become one of the major concerns of the society in the sector of public health. The death toll due to excessive consumption of alcohol is increasing exponentially. Face-to-face interaction to create awareness and to stymie the consumption of alcohol is a quixotic solution. Alcohol relapse remains a challenging problem in disorders associated with alcohol addiction which is related to spatial-temporal factors like periods and specific places. In this paper, a new generation of relapse prevention mobile application called “BeSober” is proposed. It assists users to develop abstemious habits in drinking, providing support within the intervention period, tracking risky alcoholic spots, presenting community-based support from alternative users or alcohol addiction therapists, monitoring addict's behaviour and offer personalized recommendations to assist the addict in staying sober.

Multi-label classification (MLC) is pervasive in real-world applications. Conventional MLC algorithms assume that enough ground truth labels are available for training a classifier. While in reality, obtaining ground truth labels is expensive and time-consuming. In the field of data mining, it is more efficient to use crowdsourcing for label collection. In this setting, an MLC algorithm needs to deal with the noisiness of the crowdsourced labels as well as the remaining massive unlabeled data. In this paper, we propose a deep generative model to describe the label generation process for this semi-supervised multi-label learning problem. Although deep generative models are widely used for MLC problems, no previous work could address the noisy crowdsourced multi-labels and unlabeled data simultaneously. To address this challenging problem, our novel generative model incorporates latent variables to describe the labeled/unlabeled data as well as the labeling process of crowdsourcing. We introduce an efficient sequential inference model to approximate the model posterior and infer the ground truth labels. Our experimental results on various scales of datasets demonstrate the effectiveness of our proposed model. It performs favorably against four state-of-the-art deep generative models.

Visual aesthetics of products plays an important role in the decision process when purchasing appearance-first products, e.g., clothes. Indeed, user’s aesthetic preference, which serves as a personality trait and a basic requirement, is domain independent and could be used as a bridge between domains for knowledge transfer. However, existing work has rarely considered the aesthetic information in product images for cross-domain recommendation. To this end, in this paper, we propose a new deep Aesthetic Cross-Domain Networks (ACDN), in which parameters characterizing personal aesthetic preferences are shared across networks to transfer knowledge between domains. Specifically, we first leverage an aesthetic network to extract aesthetic features. Then, we integrate these features into a cross-domain network to transfer users’ domain independent aesthetic preferences. Moreover, network cross-connections are introduced to enable dual knowledge transfer across domains. Finally, the experimental results on real-world datasets show that our proposed model ACDN outperforms benchmark methods in terms of recommendation accuracy.

The sequential recommendation, which models sequential behavioral patterns among users for the recommendation, plays a critical role in recommender systems. However, the state-of-the-art Recurrent Neural Networks (RNN) solutions rarely consider the non-linear feature interactions and non-monotone short-term sequential patterns, which are essential for user behavior modeling in sparse sequence data. In this paper, we propose a novel Recurrent Convolutional Neural Network model (RCNN). It not only utilizes the recurrent architecture of RNN to capture complex long-term dependencies, but also leverages the convolutional operation of Convolutional Neural Network (CNN) model to extract short-term sequential patterns among recurrent hidden states. Specifically, we first generate a hidden state at each time step with the recurrent layer. Then the recent hidden states are regarded as an “image”, and RCNN searches non-linear feature interactions and non-monotone local patterns via intra-step horizontal and inter-step vertical convolutional filters, respectively. Moreover, the output of convolutional filters and the hidden state are concatenated and fed into a fully-connected layer to generate the recommendation. Finally, we evaluate the proposed model using four real-world datasets from various application scenarios. The experimental results show that our model RCNN significantly outperforms the state-of-the-art approaches on sequential recommendation.

Complex Event Processing (CEP) systems extract interest situations from event flows based on event detection patterns. However, local event processing for distributed Internet of Things (IoT) has not been discussed yet. Besides, it is complex or impossible to discover such patterns in some applications of IoT. In this article, we design a complex event service to process event flows based on IoT resource models, which does not depend on existing patterns, and deals with both discrete events and continuous variables. To improve the CEP performance, local IoT resources are used for local event processing, and a lazy exchange method is designed to realize the collaborated event processing between network edges and a data center. Our evaluation shows that our solution is feasible and effective.

With the crowdsourcing of small tasks becoming easier, it is possible to obtain non-expert/imperfect labels at low cost. With low-cost imperfect labeling, it is straightforward to collect multiple labels for the same data items. This paper proposes strategies of utilizing these multiple labels for supervised learning, based on two basic ideas: majority voting and pairing. We show several interesting results based on our experiments. (i) The strategies based on the majority voting idea work well under the situation where the certainty level is high. (ii) On the contrary, the pairing strategies are more preferable under the situation where the certainty level is low. (iii) Among the majority voting strategies, soft majority voting can reduce the bias and roughness, and perform better than majority voting. (iv) Pairing can completely avoid the bias by having both sides (potentially correct and incorrect/noisy information) considered. Beta estimation is applied to reduce the impact of the noise in pairing. Our experimental results show that pairing with Beta estimation always performs well under different certainty levels. (v) All strategies investigated are labeling quality agnostic strategies for real-world applications, and some of them perform better than or at least very close to the gnostic strategies.

Traditional learning from crowdsourced labeled data consists of two stages: inferring true labels for instances from their multiple noisy labels and building a learning model using these instances with the inferred labels. This straightforward two-stage learning scheme suffers from two weaknesses: (1) the accuracy of inference may be very low; (2) useful information may be lost during inference. In this paper, we proposed a novel ensemble method for learning from crowds. Our proposed method is a meta-learning scheme. It first uses a bootstrapping process to create M sub-datasets from an original crowdsourced labeled dataset. For each sub-dataset, each instance is duplicated with different weights according to the distribution and class memberships of its multiple noisy labels. A base classifier is then trained from this extended sub-dataset. Finally, unlabeled instances are predicted by aggregating the outputs of these M base classifiers. Because the proposed method gets rid of the inference procedure and uses the full dataset to train learning models, it preserves the useful information for learning as much as possible. Experimental results on nine simulated and two real-world crowdsourcing datasets consistently show that the proposed ensemble learning method significantly outperforms five state-of-the-art methods.

With the rapid development of location-based social networks, Point-of-Interest (POI) recommendation has played an important role in helping people discover attractive locations. However, existing POI recommendation methods assume a flat structure of POIs, which are better described in a hierarchical structure in reality. Furthermore, we discover that both users' content and spatial preferences exhibit hierarchical structures. To this end, in this paper, we propose a hierarchical geographical matrix factorization model (HGMF) to utilize the hierarchical structures of both users and POIs for POI recommendation. Specifically, we first describe the POI influence degrees over regions with two-dimensional normal distribution, and learn the influence areas of different layers of POIs as the input of HGMF. Then, we perform matrix factorization on user content preference matrix, user spatial preference matrix, and POIs characteristic matrix jointly with the modeling of implicit hierarchical structures. Moreover, a two-step optimization method is proposed to learn the implicit hierarchical structure and find the solution of HGMF efficiently. Finally, we evaluate HGMF on two large-scale real-world location-based social networks datasets. Our experimental results demonstrate that it outperforms the state-of-the-art methods in terms of precision and recall.

This paper proposes a novel general label aggregation method for both binary and multi-class labeling in crowdsourcing, namely Bi-Layer Clustering (BLC), which clusters two layers of features - the conceptual-level and the physical-level features - to infer true labels of instances. BLC first clusters the instances using the conceptual-level features extracted from their multiple noisy labels and then performs clustering again using the physical-level features. It can facilitate tracking the uncertainty changes of the instances, so that the integrated labels that are likely to be falsely inferred on the conceptual layer can be easily corrected using the estimated labels on the physical layer. Experimental results on two real-world crowdsourcing data sets show that BLC outperforms seven state-of-the-art methods.

Due to low quality of crowdsourced labelers, the integrated label of each example is usually inferred from its multiple noisy labels provided by different labelers. This paper proposes a novel algorithm, Ground Truth Inference using Clustering (GTIC), to improve the quality of integrated labels for multi-class labeling. For a K labeling case, GTIC utilizes the multiple noisy label sets of examples to generate features. Then, it uses a K-Means algorithm to cluster all examples into K different groups, each of which is mapped to a specific class. Examples in the same cluster are assigned a corresponding class label. We compare GTIC with four existing multi-class ground truth inference algorithms, majority voting (MV), Dawid & Skene's (DS), ZenCrowd (ZC) and Spectral DS (SDS), on one synthetic and eight real-world datasets. Experimental results show that the performance of GTIC is significantly superior to the others in terms of both accuracy and M-AUC. Besides, the running time of GTIC is about twenty times faster than EM-based complicated inference algorithms.

Learning from imbalanced multilabel data is a challenging task. It has attracted considerable attention recently. In this paper we propose a MultiLabel Best First Over-sampling (ML-BFO) to improve the performance of multilabel classification algorithms, based on imbalance minimization and Wilson's ENN rule. Our experimental results show that ML-BFO not only duplicates fewer samples but also reduces the imbalance level much more than two state-of-the-art multilabel sampling methods, i.e., an over-sampling method LP-ROS and an under-sampling method MLeNN. Besides, ML-BFO significantly improves the performance of multilabel classification algorithms, and performs much better than LP-ROS and MLeNN.

This paper proposes a novel framework that introduces noise correction techniques to further improve label quality after ground truth inference in crowdsourcing. In the framework, an adaptive voting noise correction algorithm (AVNC) is proposed to identify and correct the most likely noises with the help of estimated qualities of labelers provided by the ground truth inference. The experimental results on two real-world datasets show that (1) the framework can improve label quality regardless of inference algorithms, especially under the circumstance that each example has a few noisy labels; and (2) since the algorithm AVNC considers both the number of and the probability of potential noises, it outperforms a baseline noise correction algorithm.

The important task of correcting label noise is addressed infrequently in literature. The difficulty of developing a robust label correction algorithm leads to this silence concerning label correction. To break the silence, we propose two algorithms to correct label noise. One utilizes self-training to re-label noise, called Self-Training Correction (STC). Another is a clustering-based method, which groups instances together to infer their ground-truth labels, called Cluster-based Correction (CC). We also adapt an algorithm from previous work, a consensus-based method called Polishing that consults with an ensemble of classifiers to change the values of attributes and labels. We simplify Polishing such that it only alters labels of instances, and call it Polishing Labels (PL). We experimentally compare our novel methods with Polishing Labels by examining their improvements on the label qualities, model qualities, and AUC metrics of binary and multi-class data sets, and ultimately conclude that only CC can significantly improve label qualities, model qualities, and AUC metrics consistently. STC and PL can improve these metrics in some cases, but not as reliably. Hence, our Cluster-based Correction method is the best.

Active learning is to select the most informative examples to request their labels. Most previous studies in active learning for multi-label classification didn't pay enough attention on label correlations. This leads to a bad performance for classification. In this paper, we proposed a chi-square statistics multi-label active learning (CSMAL) algorithm, which uses chi-square statistics to accurately evaluate correlations between labels. CSMAL considers not only positive relationships but also negative ones. It uses the average correlation between a potential label and its rest unlabeled labels as the label information for each sample-label pair. CSMAL further integrates uncertainty and label information to select example-label pairs to request labels. Our empirical results demonstrate that our proposed method CSMAL outperforms the state-of-the-art active learning methods for multi-label classification. It significantly reduces the labeling workloads and improves the performance of a classifier built.

It can be easy to collect multiple noisy labels for the same object via Internet-based crowdsourcing systems. Labelers may have bias when labeling, due to lacking expertise, dedication, and personal preference. These cause Imbalanced Multiple Noisy Labeling. In most cases, we have no information about the labeling qualities of labelers and the underlying class distributions. It is important to design agnostic solutions to utilize these noisy labels for supervised learning. We first investigate how imbalanced multiple noisy labeling affects the class distributions of training sets and the performance of classification. Then, an agnostic algorithm Positive LAbel frequency Threshold (PLAT) is proposed to deal with the imbalanced labeling issue. Simulations on eight UCI data sets with different underlying class distributions show that PLAT not only effectively deals with the imbalanced multiple noisy labeling problems that off-the-shelf agnostic methods cannot cope with, but also performs nearly the same as majority voting under the circumstances without imbalance. We also apply PLAT to eight real-world data sets with imbalanced labels collected from Amazon Mechanical Turk, and the experimental results show that PLAT is efficient and better than other ground truth inference algorithms.

Internet-based crowdsourcing systems can be viewed as a kind of loosely coupled social networks. With these systems, it is easy to collect multiple noisy labels for the same object when conducting annotation for supervised learning. Because non-expert labelers lack expertise and dedication, and have strong personal preference, they may have bias when labeling. These cause Imbalanced Multiple Noisy Labeling. In this paper, we propose an agnostic algorithm Positive LAbel frequency Threshold (PLAT) to deal with imbalanced labeling. Because of the dynamics of social networks, in most cases no information about the qualities of labelers and underlying class distributions can be acquired. PLAT does not require prior knowledge of the labeling qualities of labelers, the underlying class distributions, and the level of labeling imbalance. Simulations on eight real-world datasets with different underlying class distributions demonstrate that PLAT not only effectively deals with the imbalanced multiple noisy labeling that off-the-shelf agnostic methods cannot cope with, but also performs nearly the same as majority voting under the circumstances that labelers have no bias.

With the outsourcing of small tasks becoming easier, it is possible to obtain non-expert/imperfect labels at low cost. With low-cost imperfect labeling, it is straightforward to collect multiple labels for the same data items. This paper addresses the strategies of utilizing these multiple labels for improving the performance of supervised learning, based on two basic ideas: majority voting and pair wise solutions. We show several interesting results based on our experiments. The soft majority voting strategies can reduce the bias and roughness, and improve the performance of the directed hard majority voting strategy. Pair wise strategies can completely avoid the bias by having both sides (potential correct and incorrect/noisy information) considered (for binary classification). They have very good performance whenever there are a few or many labels available. However, it could also keep the noise. The improved variation that reduces the impact of the noisy information is recommended. All five strategies investigated are labeling quality agnostic strategies, and can be applied to real world applications directly. The experimental results show some of them perform better than or at least very close to the gnostic strategies.

This paper addresses the repeated acquisition of labels for data items when the labeling is imperfect. We examine the improvement (or lack thereof) in data quality via repeated labeling, and focus especially on the improvement of training labels for supervised induction. With the outsourcing of small tasks becoming easier, for example via Rent-A-Coder or Amazon's Mechanical Turk, it often is possible to obtain less-than-expert labeling at low cost. With low-cost labeling, preparing the unlabeled part of the data can become considerably more expensive than labeling. We present repeated-labeling strategies of increasing complexity, and show several main results. (i) Repeated-labeling can improve label quality and model quality, but not always. (ii) When labels are noisy, repeated labeling can be preferable to single labeling even in the traditional setting where labels are not particularly cheap. (iii) As soon as the cost of processing the unlabeled data is not free, even the simple strategy of labeling everything multiple times can give considerable advantage. (iv) Repeatedly labeling a carefully chosen set of points is generally preferable, and we present a robust technique that combines different notions of uncertainty to select data points for which quality should be improved. The bottom line: the results show clearly that when labeling is not perfect, selective acquisition of multiple labels is a strategy that data miners should have in their repertoire; for certain label-quality/cost regimes, the benefit is substantial.

It is often expensive to acquire data in real-world data mining applications. Most previous data mining and machine learning research, however, assumes that a fixed set of training examples is given. In this paper, we propose an online cost-sensitive framework that allows a learner to dynamically acquire examples as it learns, and to decide the ideal number of examples needed to minimize the total cost. We also propose a new strategy for Partial Example Acquisition (PAS), in which the learner can acquire examples with a subset of attribute values to reduce the data acquisition cost. Experiments on UCI datasets show that the new PAS strategy is an effective method in reducing the total cost for data acquisition.

While most software defects (i.e., bugs) are corrected and tested as part of the lengthy software development cycle, enterprise software vendors often have to release software products before all reported defects are corrected, due to deadlines and limited resources. A small number of these defects will be escalated by customers and they must be resolved immediately by the software vendors at a very high cost. In this paper, we develop an Escalation Prediction (EP) system that mines historic defect report data and predict the escalation risk of the defects for maximum net profit. More specifically, we first describe a simple and general framework to convert the maximum net profit problem to cost-sensitive learning. We then apply and compare several well-known cost-sensitive learning approaches for EP. Our experiments suggest that the cost-sensitive decision tree is the best method for producing the highest positive net profit and comprehensible results. The EP system has been deployed successfully in the product group of an enterprise software vendor.

In medical diagnosis, doctors must often determine what medical tests (e.g., X-ray and blood tests) should be ordered for a patient to minimize the total cost of medical tests and misdiagnosis. In this paper, we design cost-sensitive machine learning algorithms to model this learning and diagnosis process. Medical tests are like attributes in machine learning whose values may be obtained at a cost (attribute cost), and misdiagnoses are like misclassifications which may also incur a cost (misclassification cost). We first propose a lazy decision tree learning algorithm that minimizes the sum of attribute costs and misclassification costs. Then, we design several novel "test strategies" that can request to obtain values of unknown attributes at a cost (similar to doctors' ordering of medical tests at a cost) in order to minimize the total cost for test examples (new patients). These test strategies correspond to different situations in real-world diagnoses. We empirically evaluate these test strategies, and show that they are effective and outperform previous methods. Our results can be readily applied to real-world diagnosis tasks. A case study on heart disease is given throughout the paper