Deep hashing is widely used in large-scale image retrieval systems due to its efficient retrieval performance. However, its susceptibility to adversarial attacks limits its security in practical applications. Adversarial training is the most effective method for improving robustness, but it often leads to a significant trade-off between robustness and retrieval accuracy. In this paper, we conduct spectral analysis and find that generating high-quality hash codes requires wide-frequency response models, whereas adversarial training forces the model into spectral collapse, degrading it to a low-frequency response model and weakening its discriminability. To address this issue, we propose a Spectral-Adaptive Adversarial Hashing (SAAH) framework, which selectively preserves discriminative and task-relevant frequency components while suppressing adversarially unstable ones, enabling robust hashing without sacrificing retrieval performance. Extensive experiments on benchmark datasets demonstrate that SAAH consistently achieves a superior balance between retrieval accuracy and adversarial robustness, achieving the best performance in both retrieval accuracy and robustness compared with existing robust hashing methods.

Full-complexity machine learning models, such as the deep neural network, are non-traceable black-box, whereas the classic interpretable models, such as linear regression models, are often over-simplified, leading to lower accuracy. Model interpretability limits the application of machine learning models in management problems, which requires high prediction performance, as well as the understanding of individual features’ contributions to the model outcome. To enhance model interpretability while preserving good prediction performance, we propose a hybrid interpretable model that combines a piecewise linear component and a nonlinear component. The first component describes the explicit feature contributions by piecewise linear approximation to increase the expressiveness of the model. The other component uses a multi-layer perceptron to increase the prediction performance by capturing the high-order interactions between features and their complex nonlinear transformations. The interpretability is obtained once the model is learned in the form of shape functions for the main effects. We also provide a variant to explore the higher-order interactions among features. Experiments are conducted on synthetic and real-world datasets to demonstrate that the proposed models can achieve good interpretability by explicitly describing the main effects and the interaction effects of the features while maintaining state-of-the-art accuracy.

Topic modeling is a commonly used text analysis tool for discovering latent topics in a text corpus. However, while topics in a text corpus often exhibit a hierarchical structure (e.g., cellphone is a sub-topic of electronics), most topic modeling methods assume a flat topic structure that ignores the hierarchical dependency among topics, or utilize a predefined topic hierarchy. In this work, we present a novel Hierarchical Deep Document Model (HDDM) to learn topic hierarchies using a variational autoencoder framework. We propose a novel objective function, sum of log likelihood, instead of the widely used evidence lower bound, to facilitate the learning of hierarchical latent topic structure. The proposed objective function can directly model and optimize the hierarchical topic-word distributions at all topic levels. We conduct experiments on four real-world text datasets to evaluate the topic modeling capability of the proposed HDDM method compared to state-of-the-art hierarchical topic modeling benchmarks. Experimental results show that HDDM achieves considerable improvement over benchmarks and is capable of learning meaningful topics and topic hierarchies. To further demonstrate the practical utility of HDDM, we apply it to a real-world medical notes dataset for clinical prediction. Experimental results show that HDDM can better summarize topics in medical notes, resulting in more accurate clinical predictions.

As one of the fundamental tasks of autonomous driving, depth perception aims to perceive physical objects in three dimensions and to judge their distances away from the ego vehicle. Although great efforts have been made for depth perception, LiDAR-based and camera-based solutions have limitations with low accuracy and poor robustness for noise input. With the integration of monocular cameras and LiDAR sensors in autonomous vehicles, in this article, we introduce a two-stream architecture to learn the modality interaction representation under the guidance of an image reconstruction task to compensate for the deficiencies of each modality in a parallel manner. Specifically, in the two-stream architecture, the multi-scale cross-modality interactions are preserved via a cascading interaction network under the guidance of the reconstruction task. Next, the shared representation of modality interaction is integrated to infer the dense depth map due to the complementarity and heterogeneity of the two modalities. We evaluated the proposed solution on the KITTI dataset and CALAR synthetic dataset. Our experimental results show that learning the coupled interaction of modalities under the guidance of an auxiliary task can lead to significant performance improvements. Furthermore, our approach is competitive against the state-of-the-art models and robust against the noisy input. The source code is available at https://github.com/tonyFengye/Code/tree/master .

Learning causality from large-scale text corpora is an important task with numerous applications—for example, in finance, biology, medicine, and scientific discovery. Prior studies have focused mainly on simple causality, which only includes one cause-effect pair. However, causality is notoriously difficult to understand and analyze because of multiple cause spans and their entangled interactions. To detect complex causality, we propose a self-paced contrastive learning model, namely N2NCause, to learn entangled interactions between multiple spans. Specifically, N2NCause introduces data enhancement operations to convert implicit expressions into explicit expressions with the most rational causal connectives for the synthesis of positive samples and to invert the directed connection between a cause-effect pair for the synthesis of negative samples. To learn the semantic dependency and causal direction of positive and negative samples, self-paced contrastive learning is proposed to learn the entangled interactions among spans, including the interaction direction and interaction field. We evaluated the performance of N2NCause in three cause-effect detection tasks. The experimental results show that, with the least data annotation efforts, N2NCause demonstrates competitive performance in detecting simple cause-effect relations, and it is superior to existing solutions for the detection of complex causality.

Graph representation learning is an emerging area for graph analysis and inference. However, existing approaches for large-scale graphs either sample nodes in sequential walks or manipulate the adjacency matrices of graphs. The former approach can cause sampling bias against less-connected nodes, whereas the latter may suffer from sparsity that exists in many real-world graphs. To learn from structural information in a graph more efficiently and comprehensively, this paper proposes a new graph representation learning approach inspired by the cognitive model of spreading-activation mechanisms in human memory. This approach learns node embeddings by adopting a graph activation model that allows nodes to “activate” their neighbors and spread their own structural information to other nodes through the paths simultaneously. Comprehensive experiments demonstrate that the proposed model performs better than existing methods on several empirical datasets for multiple graph inference tasks. Meanwhile, the spreading-activation-based model is computationally more efficient than existing approaches–the training process converges after only a small number of iterations, and the training time is linear in the number of edges in a graph. The proposed method works for both homogeneous and heterogeneous graphs.

Identifying the emotional causes of mental illnesses is key to effective intervention. Existing emotion-cause analysis approaches can effectively detect simple emotion-cause expressions where only one cause and one emotion exist. However, emotions may often result from multiple causes, implicitly or explicitly, with complex interactions among these causes. Moreover, the same causes may result in multiple emotions. How to model the complex interactions between multiple emotion spans and cause spans remains under-explored. To tackle this problem, a contrastive learning-based framework is presented to detect the complex emotion-cause pairs with the introduction of negative samples and positive samples. Additionally, we developed a large-scale emotion-cause dataset with complex emotion-cause instances based on subreddits associated with mental health. Our proposed approach was compared to prevailing CNN-based, LSTM-based, Transformer-based and GNN-based methods. Extensive experiments have been conducted and the quantifiable outcomes indicate that our proposed solution achieves competitive performance on simple emotion-cause pairs and significantly outperformed baseline methods in extracting complex emotion-cause pairs. Empirical studies further demonstrated that our proposed approach can be used to reveal the emotional causes of mental disorders for effective intervention.

Adverse event detection is critical for many real-world applications including timely identification of product defects, disasters, and major socio-political incidents. In the health context, adverse drug events account for countless hospitalizations and deaths annually. Since users often begin their information seeking and reporting with online searches, examination of search query logs has emerged as an important detection channel. However, search context - including query intent and heterogeneity in user behaviors – is extremely important for extracting information from search queries, and yet the challenge of measuring and analyzing these aspects has precluded their use in prior studies. We propose DeepSAVE, a novel deep learning framework for detecting adverse events based on user search query logs. DeepSAVE uses an enriched variational autoencoder encompassing a novel query embedding and user modeling module that work in concert to address the context challenge associated with search-based detection of adverse events. Evaluation results on three large real-world event datasets show that DeepSAVE outperforms existing detection methods as well as comparison deep learning auto encoders. Ablation analysis reveals that each component of DeepSAVE significantly contributes to its overall performance. Collectively, the results demonstrate the viability of the proposed architecture for detecting adverse events from search query logs.

Online user behaviors are increasingly modulated by social media. Extant literature mainly focuses on investigating how network structures affect user behaviors. However, recent empirical results demonstrate that user behaviors and network structures usually coevolve dynamically, and topological patterns turn out to be inadequate for characterizing real-world user behaviors. In this paper, we present a dynamic model to deal with this challenge. This proposed model is mainly governed by two competing principles: homophily and homeostasis. Empirical evaluations of three online real-world datasets suggest that the proposed dynamic model can well predict long-range online user behaviors.

Bots are regarded as the most common kind of malwares in the era of Web 2.0. In recent years, Internet has been populated by hundreds of millions of bots, especially on social media. Thus, the demand on effective and efficient bot detection algorithms is more urgent than ever. Existing works have partly satisfied this requirement by way of laborious feature engineering. In this paper, we propose a deep bot detection model aiming to learn an effective representation of social user and then detect social bots by jointly modeling social behavior and content information. The proposed model learns the representation of social behavior by encoding both endogenous and exogenous factors which affect user behavior. As to the representation of content, we regard the user content as temporal text data instead of just plain text as be treated in other existing works to extract semantic information and latent temporal patterns. To the best of our knowledge, this is the first trial that applies deep learning in modeling social users and accomplishing social bot detection. Experiments on real world dataset collected from Twitter demonstrate the effectiveness of the proposed model.

In this paper, we propose a novel text representation paradigm and a set of follow-up text representation models based on cognitive psychology theories. The intuition of our study is that the knowledge implied in a large collection of documents may improve the understanding of single documents. Based on cognitive psychology theories, we propose a general text enrichment framework, study the key factors to enable activation of implicit information, and develop new text representation methods to enrich text with the implicit information. Our study aims to mimic some aspects of human cognitive procedure in which given stimulant words serve to activate understanding implicit concepts. By incorporating human cognition into text representation, the proposed models advance existing studies by mining implicit information from given text and coordinating with most existing text representation approaches at the same time, which essentially bridges the gap between explicit and implicit information. Experiments on multiple tasks show that the implicit information activated by our proposed models matches human intuition and significantly improves the performance of the text mining tasks as well.

Mining topics in short texts (e.g. tweets, instant messages) can help people grasp essential information and understand key contents, and is widely used in many applications related to social media and text analysis. The sparsity and noise of short texts often restrict the performance of traditional topic models like LDA. Recently proposed Biterm Topic Model (BTM) which models word co-occurrence patterns directly, is revealed effective for topic detection in short texts. However, BTM has two main drawbacks. It needs to manually specify topic number, which is difficult to accurately determine when facing new corpora. Besides, BTM assumes that two words in same term should belong to the same topic, which is often too strong as it does not differentiate two types of words (i.e. general words and topical words). To tackle these problems, in this paper, we propose a non-parametric topic model npCTM with the above distinction. Our model incorporates the Chinese restaurant process (CRP) into the BTM model to determine topic number automatically. Our model also distinguishes general words from topical words by jointly considering the distribution of these two word types for each word as well as word coherence information as prior knowledge. We carry out experimental studies on real-world twitter dataset. The results demonstrate the effectiveness of our method to discover coherent topics compared with the baseline methods.

Review rating prediction is of much importance for sentiment analysis and business intelligence. Existing methods work well when aspect-opinion pairs can be accurately extracted from review texts and aspect ratings are complete. The challenges of improving prediction accuracy are how to capture the semantics of review content and how to fill in the missing values of aspect ratings. In this paper, we propose a novel review rating prediction method, which improves the prediction accuracy by capturing deep semantics of review content and alleviating data missing problem of aspect ratings. The method firstly learns the latent vector representation of review content using skip-thought vectors, a state-of-the-art deep learning method, then, the missing values of aspect ratings are filled in based on users? history reviewing behaviors, finally, a novel optimization framework is proposed to predict the review rating. Experimental results on two real-world datasets demonstrate the efficacy of the proposed method.

Social media provide opportunities for policy makers to gauge pubic opinion. However, the large volumes and variety of expressions on social media have challenged traditional policy analysis and public sentiment assessment. In this article, we describe a framework for social‐media‐based public policy informatics and a system called “iMood” that addresses the needs for sentiment and network analyses of U.S. immigration and border security. iMood collects related messages on Twitter, extracts user sentiment and emotion, and constructs networks of the Twitter users, helping policy makers to identify opinion leaders, influential users, and community activists. We evaluated the sentiment, emotion, and network characteristics found in 909,035 tweets posted by over 300,000 users during three phases between May and November 2013. Statistical analyses reveal significant differences in emotion and sentiment among the 3 phases. The Twitter networks of the 3 phases also had significantly different relationship counts, network densities, and total influence scores from those of other phases. This research should contribute to developing a new framework and a new system for social‐media‐based public policy informatics, providing new empirical findings and data sets of sentiment and network analyses of U.S. immigration and border security, and demonstrating a general applicability to different domains.

The sociological theory of entrainment accounts for the synchronization of human rhythmic modalities through social interactions: they coordinate in a variety of dimensions including linguistic styles, facial expressions, music pace, applause, and so on. Though highly relevant, emotion entrainment has received little attention to date. In addition, most previous studies on entrainment are done through small scale or controlled laboratory studies. In this paper, we investigate emotion entrainment in the context of online social media. To the best of our knowledge, this is the first time that emotion entrainment has been examined on a large scale, real world setting. For this purpose, we propose a framework that can model entrainment phenomenon and measure its effect. Our framework differentiates from previous research by its model-free essential and discerning in entrainment directions. These traits enable us to model entrainment dynamics under few assumptions, and distinguish emotion flow of entrainment. In our studies, we investigate entrainment patterns under different emotion states, i.e. positive, neutral and negative. We discover that entrainments under different emotions all follow a power law distribution. Besides, people are willing to entrain to others under positive emotion, and users with positive emotion are more likely to be entrained. By inspecting the interactions between entrainment and emotion, we reveal that entrainment has an effect of negotiating different emotion types toward an even distribution.

With the rapid growth of Web 2.0, community question answering (CQA) has become a prevalent information seeking channel, in which users form interactive communities by posting questions and providing answers. Communities may evolve over time, because of changes in users' interests, activities, and new users joining the network. To better understand user interactions inCQAcommunities, it is necessary to analyze the community structures and track community evolution over time. Existing work inCQAfocuses on question searching or content quality detection, and the important problems of community extraction and evolutionary pattern detection have not been studied. In this article, we propose a probabilistic community model (PCM) to extract overlapping community structures and capture their evolution patterns inCQA. The empirical results show that our algorithm appears to improve the community extraction quality. We show empirically, using the iPhone data set, that interesting community evolution patterns can be discovered, with each evolution pattern reflecting the variation of users' interests over time. Our analysis suggests that individual users could benefit to gain comprehensive information from tracking the transition of products. We also show that the communities provide a decision‐making basis for business.

Spatio-temporal data mining is finding applications in many domains, such as public health, public safety, financial fraud detection, transportation, and product lifecycle management. Correlation analysis is an important spatio-temporal mining technique for unveiling spatial and temporal relationships among multiple event types. This paper presents a new measure for assessing and analyzing spatio-temporal cross-correlations. This measure extends Ripley’s a widely used measure of spatial correlation, with an additional temporal dimension. Empirical studies using real-world data show that the new measure can lead to a more discriminating and flexible spatio-temporal data analysis framework. In contrast with its predecessor, this measure also allows the discovery of leading (and potentially causal) event types whose occurrences precede those of other event types. Findings from analyses employing this measure may bear important managerial implications.

Despite the increased prevalence of sentiment-related information on the Web, there has been limited work on focused crawlers capable of effectively collecting not only topic-relevant but also sentiment-relevant content. In this article, we propose a novel focused crawler that incorporates topic and sentiment information as well as a graph-based tunneling mechanism for enhanced collection of opinion-rich Web content regarding a particular topic. The graph-based sentiment (GBS) crawler uses a text classifier that employs both topic and sentiment categorization modules to assess the relevance of candidate pages. This information is also used to label nodes in web graphs that are employed by the tunneling mechanism to improve collection recall. Experimental results on two test beds revealed that GBS was able to provide better precision and recall than seven comparison crawlers. Moreover, GBS was able to collect a large proportion of the relevant content after traversing far fewer pages than comparison methods. GBS outperformed comparison methods on various categories of Web pages in the test beds, including collection of blogs, Web forums, and social networking Web site content. Further analysis revealed that both the sentiment classification module and graph-based tunneling mechanism played an integral role in the overall effectiveness of the GBS crawler.

Tapping into the wisdom of the crowd, social tagging can be considered an alternative mechanism - as opposed to Web search - for organizing and discovering information on the Web. Effective tag-based recommendation of information items, such as Web resources, is a critical aspect of this social information discovery mechanism. A precise understanding of the information structure of social tagging systems lies at the core of an effective tag-based recommendation method. While most of the existing research either implicitly or explicitly assumes a simple tripartite graph structure for this purpose, we propose a comprehensive information structure to capture all types of co-occurrence information in the tagging data. Based on the proposed information structure, we further propose a unified user profiling scheme to make full use of all available information. Finally, supported by our proposed user profile, we propose a novel framework for collaborative filtering in social tagging systems. In our proposed framework, we first generate joint item-tag recommendations, with tags indicating topical interests of users in target items. These joint recommendations are then refined by the wisdom from the crowd and projected to the item space for final item recommendations. Evaluation using three real-world datasets shows that our proposed recommendation approach significantly outperformed state-of-the-art approaches.

The explosive growth of the user-generated content on the Web has offered a rich data source for mining opinions. However, the large number of diverse review sources challenges the individual users and organizations on how to use the opinion information effectively. Therefore, automated opinion mining and summarization techniques have become increasingly important. Different from previous approaches that have mostly treated product feature and opinion extraction as two independent tasks, we merge them together in a unified process by using probabilistic models. Specifically, we treat the problem of product feature and opinion extraction as a sequence labeling task and adopt Conditional Random Fields models to accomplish it. As part of our work, we develop a computational approach to construct domain specific sentiment lexicon by combining semi-structured reviews with general sentiment lexicon, which helps to identify the sentiment orientations of opinions. Experimental results on two real world datasets show that the proposed method is effective.

Abstract With the rapid development of Web 2.0, online reviews have become extremely valuable sources for mining customers' opinions. Fine‐grained opinion mining has attracted more and more attention of both applied and theoretical research. In this article, the authors study how to automatically mine product features and opinions from multiple review sources. Specifically, they propose an integration strategy to solve the issue. Within the integration strategy, the authors mine domain knowledge from semistructured reviews and then exploit the domain knowledge to assist product feature extraction and sentiment orientation identification from unstructured reviews. Finally, feature‐opinion tuples are generated. Experimental results on real‐world datasets show that the proposed approach is effective.

Accurate and timely detection of infectious disease outbreaks provides valuable information which can enable public health officials to respond to major public health threats in a timely fashion. However, disease outbreaks are often not directly observable. For surveillance systems used to detect outbreaks, noises caused by routine behavioral patterns and by special events can further complicate the detection task. Most existing detection methods combine a time series filtering procedure followed by a statistical surveillance method. The performance of this "two-step” detection method is hampered by the unrealistic assumption that the training data are outbreak-free. Moreover, existing approaches are sensitive to extreme values, which are common in real-world data sets. We considered the problem of identifying outbreak patterns in a syndrome count time series using Markov switching models. The disease outbreak states are modeled as hidden state variables which control the observed time series. A jump component is introduced to absorb sporadic extreme values that may otherwise weaken the ability to detect slow-moving disease outbreaks. Our approach outperformed several state-of-the-art detection methods in terms of detection sensitivity using both simulated and real-world data.

Abstract User‐generated content on the Web has become an extremely valuable source for mining and analyzing user opinions on any topic. Recent years have seen an increasing body of work investigating methods to recognize favorable and unfavorable sentiments toward specific subjects from online text. However, most of these efforts focus on English and there have been very few studies on sentiment analysis of Chinese content. This paper aims to address the unique challenges posed by Chinese sentiment analysis. We propose a rule‐based approach including two phases: (1) determining each sentence's sentiment based on word dependency, and (2) aggregating sentences to predict the document sentiment. We report the results of an experimental study comparing our approach with three machine learning‐based approaches using two sets of Chinese articles. These results illustrate the effectiveness of our proposed method and its advantages against learning‐based approaches.

The 12 papers in this special section focus on intelligence and security informatics. They are summarized here.

Recommender systems are being widely applied in many application settings to suggest products, services, and information items to potential consumers. Collaborative filtering, the most successful recommendation approach, makes recommendations based on past transactions and feedback from consumers sharing similar interests. A major problem limiting the usefulness of collaborative filtering is the sparsity problem, which refers to a situation in which transactional or feedback data is sparse and insufficient to identify similarities in consumer interests. In this article, we propose to deal with this sparsity problem by applying an associative retrieval framework and related spreading activation algorithms to explore transitive associations among consumers through their past transactions and feedback. Such transitive associations are a valuable source of information to help infer consumer interests and can be explored to deal with the sparsity problem. To evaluate the effectiveness of our approach, we have conducted an experimental study using a data set from an online bookstore. We experimented with three spreading activation algorithms including a constrained Leaky Capacitor algorithm, a branch-and-bound serial symbolic search algorithm, and a Hopfield net parallel relaxation search algorithm. These algorithms were compared with several collaborative filtering approaches that do not consider the transitive associations: a simple graph search approach, two variations of the user-based approach, and an item-based approach. Our experimental results indicate that spreading activation-based approaches significantly outperformed the other collaborative filtering methods as measured by recommendation precision, recall, the F-measure, and the rank score. We also observed the over-activation effect of the spreading activation approach, that is, incorporating transitive associations with past transactional data that is not sparse may "dilute" the data used to infer user preferences and lead to degradation in recommendation performance.

Abstract It has become increasingly difficult to locate relevant information on the Web, even with the help of Web search engines. Two approaches to addressing the low precision and poor presentation of search results of current search tools are studied: meta‐search and document categorization. Meta‐search engines improve precision by selecting and integrating search results from generic or domain‐specific Web search engines or other resources. Document categorization promises better organization and presentation of retrieved results. This article introduces MetaSpider, a meta‐search engine that has real‐time indexing and categorizing functions. We report in this paper the major components of MetaSpider and discuss related technical approaches. Initial results of a user evaluation study comparing MetaSpider, NorthernLight, and MetaCrawler in terms of clustering performance and of time and effort expended show that MetaSpider performed best in precision rate, but disclose no statistically significant differences in recall rate and time requirements. Our experimental study also reveals that MetaSpider exhibited a higher level of automation than the other two systems and facilitated efficient searching by providing the user with an organized, comprehensive view of the retrieved documents.

We are investigating techniques for developing distributed and adaptive collections of information agents that coordinate to retrieve, filter and fuse information relevant to the user, task and situation, as well as anticipate user's information needs. In our system of agents, information gathering is seamlessly integrated with decision support. The task for which particular information is requested of the agents does not remain in the user's head but it is explicitly represented and supported through agent collaboration. In this paper we present the distributed system architecture, agent collaboration interactions, and a reusable set of software components for structuring agents. The system architecture has three types of agents: Interface agents interact with the user receiving user specifications and delivering results. They acquire, model, and utilize user preferences to guide system coordination in support of the user's tasks. Task agents help users perform tasks by formulating problem solving plans and carrying out these plans through querying and exchanging information with other software agents. Information agents provide intelligent access to a heterogeneous collection of information sources. We have implemented this system framework and are developing collaborating agents in diverse complex real world tasks, such as organizational decision making, investment counseling, health care and electronic commerce.