Search

Work on social media rumour verification utilises signals from posts, their propagation and users involved. Other lines of work target identifying and fact-checking claims based on information from Wikipedia, or trustworthy news articles without considering social media context. However works combining the information from social media with external evidence from the wider web are lacking. To facilitate research in this direction, we release a novel dataset, PHEMEPlus, an extension of the PHEME benchmark, which contains social media conversations as well as relevant external evidence for each rumour. We demonstrate the effectiveness of incorporating such evidence in improving rumour verification models. Additionally, as part of the evidence collection, we evaluate various ways of query formulation to identify the most effective method., 10 pages, 1 figure, 5 tables, presented in the Fifth Fact Extraction and VERification Workshop (FEVER). 2022

With the increasing popularity of posting multimodal messages online, many recent studies have been carried out utilizing both textual and visual information for multi-modal sarcasm detection. In this paper, we investigate multimodal sarcasm detection from a novel perspective by constructing a cross-modal graph for each instance to explicitly draw the ironic relations between textual and visual modalities. Specifically, we first detect the objects paired with descriptions of the image modality, enabling the learning of important visual information. Then, the descriptions of the objects are served as a bridge to determine the importance of the association between the objects of image modality and the contextual words of text modality, so as to build a cross-modal graph for each multi-modal instance. Furthermore, we devise a cross-modal graph convolutional network to make sense of the incongruity relations between modalities for multi-modal sarcasm detection. Extensive experimental results and in-depth analysis show that our model achieves state-of-the-art performance in multi-modal sarcasm detection.

Zero-shot stance detection (ZSSD) aims to detect the stance for an unseen target during the inference stage. In this paper, we propose a joint contrastive learning (JointCL) framework, which consists of stance contrastive learning and target-aware prototypical graph contrastive learning. Specifically, a stance contrastive learning strategy is employed to better generalize stance features for unseen targets. Further, we build a prototypical graph for each instance to learn the target-based representation, in which the prototypes are deployed as a bridge to share the graph structures between the known targets and the unseen ones. Then a novel target-aware prototypical graph contrastive learning strategy is devised to generalize the reasoning ability of target-based stance representations to the unseen targets. Extensive experiments on three benchmark datasets show that the proposed approach achieves state-ofthe- art performance in the ZSSD task.

This paper presents an unsupervised framework for jointly modeling topic content and discourse behavior in microblog conversations. Concretely, we propose a neural model to discover word clusters indicating what a conversation concerns (i.e., topics) and those reflecting how participants voice their opinions (i.e., discourse). Extensive experiments show that our model can yield both coherent topics and meaningful discourse behavior. Further study shows that our topic and discourse representations can benefit the classification of microblog messages, especially when they are jointly trained with the classifier., Accepted in Transactions of the Association for Computational Linguistics

In recent years participatory budgeting (PB) in Scotland has grown from a handful of community-led processes to a movement supported by local and national government. This is epitomized by an agreement between the Scottish Government and the Convention of Scottish Local Authorities (COSLA) that at least 1% of local authority budgets will be subject to PB. This ongoing research paper explores the challenges that emerge from this 'scaling up' or 'mainstreaming' across the 32 local authorities that make up Scotland. The main objective is to evaluate local authority use of the digital platform Consul, which applies Natural Language Processing (NLP) to address these challenges. This project adopts a qualitative longitudinal design with interviews, observations of PB processes, and analysis of the digital platform data. Thematic analysis is employed to capture the major issues and themes which emerge. Longitudinal analysis then explores how these evolve over time. The potential for 32 live study sites provides a unique opportunity to explore discrete political and social contexts which materialize and allow for a deeper dive into the challenges and issues that may exist, something a wider cross-sectional study would miss. Initial results show that issues and challenges which come from scaling up may be tackled using NLP technology which, in a previous controlled use case-based evaluation, has shown to improve the effectiveness of citizen participation., 7 pages, presented at the 14th International Conference on Theory and Practice of Electronic Governance 2021. arXiv admin note: text overlap with arXiv:2109.09517

As the digital news industry becomes the main channel of information dissemination, the adverse impact of fake news is explosively magnified. The credibility of a news report should not be considered in isolation. Rather, previously published news articles on the similar event could be used to assess the credibility of a news report. Inspired by this, we propose a BERT-based multimodal unreliable news detection framework, which captures both textual and visual information from unreliable articles utilising the contrastive learning strategy. The contrastive learner interacts with the unreliable news classifier to push similar credible news (or similar unreliable news) closer while moving news articles with similar content but opposite credibility labels away from each other in the multimodal embedding space. Experimental results on a COVID-19 related dataset, ReCOVery, show that our model outperforms a number of competitive baseline in unreliable news detection., Accepted by the CIKM 2021

Understanding patient opinions expressed towards healthcare services in online platforms could allow healthcare professionals to respond to address patients' concerns in a timely manner. Extracting patient opinion towards various aspects of health services is closely related to aspect-based sentiment analysis (ABSA) in which we need to identify both opinion targets and target-specific opinion expressions. The lack of aspect-level annotations however makes it difficult to build such an ABSA system. This paper proposes a joint learning framework for simultaneous unsupervised aspect extraction at the sentence level and supervised sentiment classification at the document level. It achieves 98.2% sentiment classification accuracy when tested on the reviews about healthcare services collected from Yelp, outperforming several strong baselines. Moreover, our model can extract coherent aspects and can automatically infer the distribution of aspects under different polarities without requiring aspect-level annotations for model learning.

In this paper, we propose the Brand-Topic Model (BTM) which aims to detect brand-associated polarity-bearing topics from product reviews. Different from existing models for sentiment-topic extraction which assume topics are grouped under discrete sentiment categories such as ‘positive’, ‘negative’ and ‘neural’, BTM is able to automatically infer real-valued brand-associated sentiment scores and generate fine-grained sentiment-topics in which we can observe continuous changes of words under a certain topic (e.g., ‘shaver’ or ‘cream’) while its associated sentiment gradually varies from negative to positive. BTM is built on the Poisson factorisation model with the incorporation of adversarial learning. It has been evaluated on a dataset constructed from Amazon reviews. Experimental results show that BTM outperforms a number of competitive baselines in brand ranking, achieving a better balance of topic coherence and unique-ness, and extracting better-separated polarity-bearing topics.

The development of democratic systems is a crucial task as confirmed by its selection as one of the Millennium Sustainable Development Goals by the United Nations. In this article, we report on the progress of a project that aims to address barriers, one of which is information overload, to achieving effective direct citizen participation in democratic decision-making processes. The main objectives are to explore if the application of Natural Language Processing (NLP) and machine learning can improve citizens' experience of digital citizen participation platforms. Taking as a case study the "Decide Madrid" Consul platform, which enables citizens to post proposals for policies they would like to see adopted by the city council, we used NLP and machine learning to provide new ways to (a) suggest to citizens proposals they might wish to support; (b) group citizens by interests so that they can more easily interact with each other; (c) summarise comments posted in response to proposals; (d) assist citizens in aggregating and developing proposals. Evaluation of the results confirms that NLP and machine learning have a role to play in addressing some of the barriers users of platforms such as Consul currently experience., Comment: 19 pages, 5 figures, 4 tables, to appear in Digital Government: Research and Practice (DGOV)

Detecting sarcastic expressions could promote the understanding of natural language in social media. In this paper, we revisit sarcasm detection from a novel perspective, so as to account for the long-range literal sentiment inconsistencies. More concretely, we explore a novel scenario of constructing an affective graph and a dependency graph for each sentence based on the affective information retrieved from external affective commonsense knowledge and the syntactical information of the sentence. Based on it, an Affective Dependency Graph Convolutional Network (ADGCN) framework is proposed to draw long-range incongruity patterns and inconsistent expressions over the context for sarcasm detection by means with interactively modeling the affective and dependency information. Experimental results on multiple benchmark datasets show that our proposed approach outperforms the current state-of-the-art methods in sarcasm detection.

Target plays an essential role in stance detection of an opinionated review/claim, since the stance expressed in the text often depends on the target. In practice, we need to deal with targets unseen in the annotated training data. As such, detecting stance for an unknown or unseen target is an important research problem. This paper presents a novel approach that automatically identifies and adapts the target-dependent and target-independent roles that a word plays with respect to a specific target in stance expressions, so as to achieve cross-target stance detection. More concretely, we explore a novel solution of constructing heterogeneous target-adaptive pragmatics dependency graphs (TPDG) for each sentence towards a given target. An in-target graph is constructed to produce inherent pragmatics dependencies of words for a distinct target. In addition, another cross-target graph is constructed to develop the versatility of words across all targets for boosting the learning of dominant word-level stance expressions available to an unknown target. A novel graph-aware model with interactive Graphical Convolutional Network (GCN) blocks is developed to derive the target-adaptive graph representation of the context for stance detection. The experimental results on a number of benchmark datasets show that our proposed model outperforms state-of-the-art methods in cross-target stance detection.\ud \ud

The flexibility of the inference process in Variational Autoencoders (VAEs) has recently led to revising traditional probabilistic topic models giving rise to Neural Topic Models (NTMs). Although these approaches have achieved significant results, surprisingly very little work has been done on how to disentangle the latent topics. Existing topic models when applied to reviews may extract topics associated with writers' subjective opinions mixed with those related to factual descriptions such as plot summaries in movie and book reviews. It is thus desirable to automatically separate opinion topics from plot/neutral ones enabling a better interpretability. In this paper, we propose a neural topic model combined with adversarial training to disentangle opinion topics from plot and neutral ones. We conduct an extensive experimental assessment introducing a new collection of movie and book reviews paired with their plots, namely MOBO dataset, showing an improved coherence and variety of topics, a consistent disentanglement rate, and sentiment classification performance superior to other supervised topic models., Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL 2021

Neural attention mechanism has achieved many successes in various tasks in natural language processing. However, existing neural attention models based on a densely connected network are loosely related to the attention mechanism found in psychology and neuroscience. Motivated by the finding in neuroscience that human possesses the template-searching attention mechanism, we propose to use convolution operation to simulate attentions and give a mathematical explanation of our neural attention model. We then introduce a new network architecture, which combines a recurrent neural network with our convolution-based attention model and further stacks an attention-based neural model to build a hierarchical sentiment classification model. The experimental results show that our proposed models can capture salient parts of the text to improve the performance of sentiment classification at both the sentence level and the document level.

Existing methods for question answering over knowledge bases (KBQA) ignore the consideration of the model prediction uncertainties. We argue that estimating such uncertainties is crucial for the reliability and interpretability of KBQA systems. Therefore, we propose a novel end-to-end KBQA model based on Bayesian Neural Network (BNN) to estimate uncertainties arose from both model and data. To our best knowledge, we are the first to consider the uncertainty estimation problem for the KBQA task using BNN. The proposed end-to-end model integrates entity detection and relation prediction into a unified framework, and employs BNN to model entity and relation under the given question semantics, transforming network weights into distributions. Therefore, predictive distributions can be estimated by sampling weights and forward inputs through the network multiple times. Uncertainties can be further quantified by calculating the variances of predictive distributions. The experimental results demonstrate the effectiveness of uncertainties in both the misclassification detection task and cause of error detection task. Furthermore, the proposed model also achieves comparable performance compared to the existing state-of-the-art approaches on SimpleQuestions dataset.

With the rapid growth of social media in the past decade, the news are no longer controlled by just a few mainstream sources. Users themselves create large numbers of potentially fictitious rumours, necessitating automated veracity classification systems. Here we present a novel approach towards automatically classifying rumours circulating on Twitter with respect to their veracity. We use a model built on Variational Autoencoder which disentangles the informational content of a tweet from the manner in which the information is written. This is achieved by obtaining latent topic vectors in an adversarial learning setting using the auxiliary task of stance classification. The latent vectors learnt in this way are used to predict rumour veracity, obtaining state-of-the-art accuracy scores on the PHEME dataset.

Social media has become the new playground for bullies. Young people are now regularly exposed to a wide range of abuse online. In response to the increasing prevalence of cyberbullying, online social networks have increased efforts to clamp down on online abuse but unfortunately, the nature, complexity and sheer volume of cyberbullying means that many cyberbullying incidents go undetected. BullStop is a mobile app for detecting and preventing cyberbullying and online abuse on social media platforms. It uses deep learning models to identify instances of cyberbullying and can automatically initiate actions such as deleting offensive messages and blocking bullies on behalf of the user. Our system not only achieves impressive prediction results but also demonstrates excellent potential for use in real-world scenarios and is freely available on the Google Play Store.

We introduce CHIME, a cross-passage hierarchical memory network for question answering (QA) via text generation. It extends XLNet introducing an auxiliary memory module consisting of two components: the context memory collecting cross-passage evidences, and the answer memory working as a buffer continually refining the generated answers. Empirically, we show the efficacy of the proposed architecture in the multi-passage generative QA, outperforming the state-of-the-art baselines with better syntactically well-formed answers and increased precision in addressing the questions of the AmazonQA review dataset. An additional qualitative analysis revealed the interpretability introduced by the memory module., COLING 2020

In most previous studies, information about aspects in sentences is considered important for the Aspect-based Sentiment Analysis (ABSA task and therefore various attention mechanisms have been explored to leverage interactions between aspects and context. However, some sentiment expressions carry the same polarity regardless of the aspects they are associated with. In such cases, it is not necessary to incorporate aspect information for ABSA. In fact, in our experiments, we find that blindly leveraging interactions between aspects and context as features may introduce noise when analyzing those aspect-invariant sentiment expressions, especially when facing with limited aspect-related annotated data. Hence, in this paper, we propose an Adversarial Multi-task Learning Framework to identify the aspect-invariant/dependent sentiment expressions automatically without requiring extra annotations. In addition, we use a gating mechanism to control the contribution of representations derived from aspect-invariant and aspect-dependent hidden states when generating the final contextual sentiment representations for the given aspect. This essentially allows the exploitation of aspect-invariant sentiment features for better ABSA results. Experimental results on two benchmark datasets show that extending existing neural models using our proposed framework achieves superior performance. In addition, the aspect-invariant data extracted by our framework can be considered as pivot features for better transfer learning of the ABSA models on unseen aspects.&nbsp

Climate models (CM) are used to evaluate the impact of climate change on the risk of floods and heavy precipitation events. However, these numerical simulators produce outputs with low spatial resolution that exhibit difficulties representing precipitation events accurately. This is mainly due to computational limitations on the spatial resolution used when simulating multi-scale weather dynamics in the atmosphere. To improve the prediction of high resolution precipitation we apply a Deep Learning (DL) approach using input data from a reanalysis product, that is comparable to a climate model’s output, but can be directly related to precipitation observations at a given time and location. Further, our input excludes local precipitation, but includes model fields (weather variables) that are more predictable and generalizable than local precipitation. To this end, we present TRU-NET (Temporal Recurrent U-Net), an encoder-decoder model featuring a novel 2D cross attention mechanism between contiguous convolutional-recurrent layers to effectively model multi-scale spatio-temporal weather processes. We also propose a non-stochastic variant of the conditional-continuous (CC) loss function to capture the zero-skewed patterns of rainfall. Experiments show that our models, trained with our CC loss, consistently attain lower RMSE and MAE scores than a DL model prevalent in precipitation downscaling and outperform a state-of-the-art dynamical weather model. Moreover, by evaluating the performance of our model under various data formulation strategies, for the training and test sets, we show that there is enough data for our deep learning approach to output robust, high-quality results across seasons and varying regions.

Recently, advances in neural network approaches have achieved many successes in both sentiment classification and probabilistic topic modelling. On the one hand, latent topics derived from the global context of documents could be helpful in capturing more accurate word semantics and hence could potentially improve the sentiment classification accuracy. On the other hand, the word-level attention vectors obtained during the learning of sentiment classifiers could carry word-level polarity information and can be used to guide the discovery of topics in topic modelling. This paper proposes a multi-task learning framework which jointly learns a sentiment classifier and a topic model by making the word-level latent topic distributions in the topic model to be similar to the word-level attention vectors in the classifier through mutual learning. Experimental results on the Yelp and IMDB datasets verify the superior performance of the proposed framework over strong baselines on both sentiment classification accuracy and topic modelling evaluation results including perplexity and topic coherence measures. The proposed framework also extracts more interpretable topics compared to other conventional topic models and neural topic models.

We address and formalise the task of sequence-to-sequence (seq2seq) cross-modal retrieval. Given a sequence of text passages as query, the goal is to retrieve a sequence of images that best describes and aligns with the query. This new task extends the traditional cross-modal retrieval, where each image-text pair is treated independently ignoring broader context. We propose a novel variational recurrent seq2seq (VRSS) retrieval model for this seq2seq task. Unlike most cross-modal methods, we generate an image vector corresponding to the latent topic obtained from combining the text semantics and context. This synthetic image embedding point associated with every text embedding point can then be employed for either image generation or image retrieval as desired. We evaluate the model for the application of stepwise illustration of recipes, where a sequence of relevant images are retrieved to best match the steps described in the text. To this end, we build and release a new Stepwise Recipe dataset for research purposes, containing 10K recipes (sequences of image-text pairs) having a total of 67K image-text pairs. To our knowledge, it is the first publicly available dataset to offer rich semantic descriptions in a focused category such as food or recipes. Our model is shown to outperform several competitive and relevant baselines in the experiments. We also provide qualitative analysis of how semantically meaningful the results produced by our model are through human evaluation and comparison with relevant existing methods.

Research into cyberbullying detection has increased in recent years, due in part to the proliferation of cyberbullying across social media and its detrimental effect on young people. A growing body of work is emerging on automated approaches to cyberbullying detection. These approaches utilise machine learning and natural language processing techniques to identify the characteristics of a cyberbullying exchange and automatically detect cyberbullying by matching textual data to the identified traits. In this paper, we present a systematic review of published research (as identified via Scopus, ACM and IEEE Xplore bibliographic databases) on cyberbullying detection approaches. On the basis of our extensive literature review, we categorise existing approaches into 4 main classes, namely; supervised learning, lexicon based, rule based and mixed-initiative approaches. Supervised learning-based approaches typically use classifiers such as SVM and Naïve Bayes to develop predictive models for cyberbullying detection. Lexicon based systems utilise word lists and use the presence of words within the lists to detect cyberbullying. Rules-based approaches match text to predefined rules to identify bullying and mixed-initiatives approaches combine human-based reasoning with one or more of the aforementioned approaches. We found lack of quality representative labelled datasets and non-holistic consideration of cyberbullying by researchers when developing detection systems are two key challenges facing cyberbullying detection research. This paper essentially maps out the state-of-the-art in cyberbullying detection research and serves as a resource for researchers to determine where to best direct their future research efforts in this field.

Social media platforms allow users to express their opinions towards various topics online. Oftentimes, users’ opinions are not static, but might be changed over time due to the influences from their neighbors in social networks or updated based on arguments encountered that undermine their beliefs. In this paper, we propose to use a Recurrent Neural Network (RNN) to model each user’s posting behaviors on Twitter and incorporate their neighbors’ topic-associated context as attention signals using an attention mechanism for user-level stance prediction. Moreover, our proposed model operates in an online setting in that its parameters are continuously updated with the Twitter stream data and can be used to predict user’s topic-dependent stance. Detailed evaluation on two Twitter datasets, related to Brexit and US General Election, justifies the superior performance of our neural opinion dynamics model over both static and dynamic alternatives for user-level stance prediction.

In our world with full of uncertainty, debates and argumentation contribute to the progress of science and society. Despite of the increasing attention to characterize human arguments, most progress made so far focus on the debate outcome, largely ignoring the dynamic patterns in argumentation processes. This paper presents a study that automatically analyzes the key factors in argument persuasiveness, beyond simply predicting who will persuade whom. Specifically, we propose a novel neural model that is able to dynamically track the changes of latent topics and discourse in argumentative conversations, allowing the investigation of their roles in influencing the outcomes of persuasion. Extensive experiments have been conducted on argumentative conversations on both social media and supreme court. The results show that our model outperforms state-of-the-art models in identifying persuasive arguments via explicitly exploring dynamic factors of topic and discourse. We further analyze the effects of topics and discourse on persuasiveness, and find that they are both useful - topics provide concrete evidence while superior discourse styles may bias participants, especially in social media arguments. In addition, we draw some findings from our empirical results, which will help people better engage in future persuasive conversations., 12 pages, accepted by The Web Conference 2020 (WWW'20)

Background Aedes albopictus is a major vector for several tropical infectious diseases. Characterization of Ae. albopictus development under natural conditions is crucial for monitoring vector population expansion, dengue virus transmission, and disease outbreak preparedness. Methods This study employed mosquito traits as a proxy to understanding life-table traits in mosquitoes using a semi–field study. Ae. albopictus larval and adult life-table experiments were conducted using microcosms under semi-field conditions in Guangzhou. Stage-specific development times and survivorship rates were determined and compared under semi-field conditions in different seasons from early summer (June) to winter (January), to determine the lower temperature limit for larval development and adult survivorship and reproductivity. Results The average egg- hatching rate was 60.1%, with the highest recorded in October (77.1%; mid-autumn). The larval development time was on average 13.2 days (range, 8.5–24.1 days), with the shortest time observed in September(8.7 days; early autumn) and longest in November (22.8 days). The pupation rates of Ae. albopictus larvae were on average 88.9% (range, 81.6–93.4%); they were stable from June to September but decreased from October to November. The adult emergence rates were on average 82.5% (range, 76.8–87.9%) and decreased from July to November. The median survival time of Ae. albopictus adults was on average 7.4 (range, 4.5–9.8), with the shortest time recorded in September. The average lifetime egg mass under semi-field conditions was 37.84 eggs/female. The larvae could develop into adults at temperatures as low as 12.3°C, and the adults could survive for 30.0 days at 16.3°C and still produce eggs. Overall, correlation analysis found that mean temperature and relative humidity were variables significantly affecting larval development and adult survivorship. Conclusion Ae. albopictus larvae could develop and emerge and the adults could survive and produce eggs in early winter in Guangzhou. The major impact of changes in ambient temperature, relative humidity, and light intensity was on the egg hatching rates, adult survival time, and egg mass production, rather than on pupation or adult emergence rates.

To extract the structured representations of open-domain events, Bayesian graphical models have made some progress. However, these approaches typically assume that all words in a document are generated from a single event. While this may be true for short text such as tweets, such an assumption does not generally hold for long text such as news articles. Moreover, Bayesian graphical models often rely on Gibbs sampling for parameter inference which may take long time to converge. To address these limitations, we propose an event extraction model based on Generative Adversarial Nets, called Adversarial-neural Event Model (AEM). AEM models an event with a Dirichlet prior and uses a generator network to capture the patterns underlying latent events. A discriminator is used to distinguish documents reconstructed from the latent events and the original documents. A byproduct of the discriminator is that the features generated by the learned discriminator network allow the visualization of the extracted events. Our model has been evaluated on two Twitter datasets and a news article dataset. Experimental results show that our model outperforms the baseline approaches on all the datasets, with more significant improvements observed on the news article dataset where an increase of 15\% is observed in F-measure., Accepted by EMNLP-IJCNLP 2019

Multiple emotions with different intensities are often evoked by events described in documents. Oftentimes, such event information is hidden and needs to be discovered from texts. Unveiling the hidden event information can help to understand how the emotions are evoked and provide explainable results. However, existing studies often ignore the latent event information. In this paper, we proposed a novel interpretable relevant emotion ranking model with the event information incorporated into a deep learning architecture using the event-driven attentions. Moreover, corpus-level event embeddings and document-level event distributions are introduced respectively\ud to consider the global events in corpus and the document-specific events simultaneously. Experimental results on three real-world corpora show that the proposed approach performs remarkably better than the state-of-the-art emotion detection approaches and multi-label approaches. Moreover, interpretable results can be obtained to shed light on the events which trigger certain emotions.

Online reviews have become an important source of information for users before making an informed purchase decision. Early reviews of a product tend to have a high impact on the subsequent product sales. In this paper, we take the initiative to study the behavior characteristics of early reviewers through their posted reviews on two real-world large e-commerce platforms, i.e., Amazon and Yelp. In specific, we divide product lifetime into three consecutive stages, namely early , majority , and laggards . A user who has posted a review in the early stage is considered as an early reviewer. We quantitatively characterize early reviewers based on their rating behaviors, the helpfulness scores received from others and the correlation of their reviews with product popularity. We have found that (1) an early reviewer tends to assign a higher average rating score; and (2) an early reviewer tends to post more helpful reviews. Our analysis of product reviews also indicates that early reviewers’ ratings and their received helpfulness scores are likely to influence product popularity. By viewing the review posting process as a multiplayer competition game, we propose a novel margin-based embedding model for early reviewer prediction. Extensive experiments on two different e-commerce datasets have shown that our proposed approach outperforms a number of competitive baselines.

Topic models are widely used for thematic structure discovery in text. But traditional topic models often require dedicated inference procedures for specific tasks at hand. Also, they are not designed to generate word-level semantic representations. To address these limitations, we propose a topic modeling approach based on Generative Adversarial Nets (GANs), called Adversarial-neural Topic Model (ATM). The proposed ATM models topics with Dirichlet prior and employs a generator network to capture the semantic patterns among latent topics. Meanwhile, the generator could also produce word-level semantic representations. To illustrate the feasibility of porting ATM to tasks other than topic modeling, we apply ATM for open domain event extraction. Our experimental results on the two public corpora show that ATM generates more coherence topics, outperforming a number of competitive baselines. Moreover, ATM is able to extract meaningful events from news articles., Published at the journal Information Processing & Management

Objective\ud \ud A drug–drug interaction (DDI) is a situation in which a drug affects the activity of another drug synergistically or antagonistically when being administered together. The information of DDIs is crucial for healthcare professionals to prevent adverse drug events. Although some known DDIs can be found in purposely-built databases such as DrugBank, most information is still buried in scientific publications. Therefore, automatically extracting DDIs from biomedical texts is sorely needed.\ud \ud Methods and material\ud \ud In this paper, we propose a novel position-aware deep multi-task learning approach for extracting DDIs from biomedical texts. In particular, sentences are represented as a sequence of word embeddings and position embeddings. An attention-based bidirectional long short-term memory (BiLSTM) network is used to encode each sentence. The relative position information of words with the target drugs in text is combined with the hidden states of BiLSTM to generate the position-aware attention weights. Moreover, the tasks of predicting whether or not two drugs interact with each other and further distinguishing the types of interactions are learned jointly in multi-task learning framework.\ud \ud Results\ud \ud The proposed approach has been evaluated on the DDIExtraction challenge 2013 corpus and the results show that with the position-aware attention only, our proposed approach outperforms the state-of-the-art method by 0.99% for binary DDI classification, and with both position-aware attention and multi-task learning, our approach achieves a micro F-score of 72.99% on interaction type identification, outperforming the state-of-the-art approach by 1.51%, which demonstrates the effectiveness of the proposed approach.

Online Judge (OJ) systems have been widely used in many areas, including programming, mathematical problems solving, and job interviews. Unlike other online learning systems, such as Massive Open Online Course, most OJ systems are designed for self-directed learning without the intervention of teachers. Also, in most OJ systems, problems are simply listed in volumes and there is no clear organization of them by topics or difficulty levels. As such, problems in the same volume are mixed in terms of topics or difficulty levels. By analyzing large-scale users’ learning traces, we observe that there are two major learning modes (or patterns). Users either practice problems in a sequential manner from the same volume regardless of their topics or they attempt problems about the same topic, which may spread across multiple volumes. Our observation is consistent with the findings in classic educational psychology. Based on our observation, we propose a novel two-mode Markov topic model to automatically detect the topics of online problems by jointly characterizing the two learning modes. For further predicting the difficulty level of online problems, we propose a competition-based expertise model using the learned topic information. Extensive experiments on three large OJ datasets have demonstrated the effectiveness of our approach in three different tasks, including skill topic extraction, expertise competition prediction and problem recommendation.

Emotion cause extraction aims to identify the reasons behind a certain emotion expressed in text. It is a much more difficult task compared to emotion classification. Inspired by recent advances in using deep memory networks for question answering (QA), we propose a new approach which considers emotion cause identification as a reading comprehension task in QA. Inspired by convolutional neural networks, we propose a new mechanism to store relevant context in different memory slots to model context information. Our proposed approach can extract both word level sequence features and lexical features. Performance evaluation shows that our method achieves the state-of-the-art performance on a recently released emotion cause dataset, outperforming a number of competitive baselines by at least 3.01% in F-measure.

Stance classification, which aims at detecting the stance expressed in text towards a specific target, is an emerging problem in sentiment analysis. A major difference between stance classification and traditional aspect-level sentiment classification is that the identification of stance is dependent on target which might not be explicitly mentioned in text. This indicates that apart from text content, the target information is important to stance detection. To this end, we propose a neural network-based model, which incorporates target-specific information into stance classification by following a novel attention mechanism. In specific, the attention mechanism is expected to locate the critical parts of text which are related to target. Our evaluations on both the English and Chinese Stance Detection datasets show that the proposed model achieves the state-of-the-art performance.