Search

Hate speech recognizers (HSRs) can be the panacea for containing hate in social media or can result in the biggest form of prejudice-based censorship hindering people to express their true selves. In this paper, we hypothesized how massive use of syntax can reduce the prejudice effect in HSRs. To explore this hypothesis, we propose Unintended-bias Visualizer based on Kermit modeling (KERM-HATE): a syntax-based HSR, which is endowed with syntax heat parse trees used as a post-hoc explanation of classifications. KERM-HATE significantly outperforms BERT-based, RoBERTa-based and XLNet-based HSR on standard datasets. Surprisingly this result is not sufficient. In fact, the post-hoc analysis on novel datasets on recent divisive topics shows that even KERM-HATE carries the prejudice distilled from the initial corpus. Therefore, although tests on standard datasets may show higher performance, syntax alone cannot drive the “attention” of HSRs to ethically-unbiased features.

Readability is an active field of research in the late nineteenth century and vigorously persuaded to date. The recent boom in data-driven machine learning has created a viable path forward for readability classification and ranking. The evaluation of text readability is a time-honoured issue with even more relevance in today’s information-rich world. This paper addresses the task of readability assessment for the English language. Given the input sentences, the objective is to predict its level of readability, which corresponds to the level of literacy anticipated from the target readers. This readability aspect plays a crucial role in drafting and comprehending processes of English language learning. Selecting and presenting a suitable collection of sentences for English Language Learners may play a vital role in enhancing their learning curve. In this research, we have used 30,000 English sentences for experimentation. Additionally, they have been annotated into seven different readability levels using Flesch Kincaid. Later, various experiments were conducted using five Machine Learning algorithms, i.e., KNN, SVM, LR, NB, and ANN. The classification models render excellent and stable results. The ANN model obtained an F-score of 0.95% on the test set. The developed model may be used in education setup for tasks such as language learning, assessing the reading and writing abilities of a learner.

Customer satisfaction and their positive sentiments are some of the various goals for successful companies. However, analyzing customer reviews to predict accurate sentiments have been proven to be challenging and time-consuming due to high volumes of collected data from various sources. Several researchers approach this with algorithms, methods, and models. These include machine learning and deep learning (DL) methods, unigram and skip-gram based algorithms, as well as the Artificial Neural Network (ANN) and bag-of-word (BOW) regression model. Studies and research have revealed incoherence in polarity, model overfitting and performance issues, as well as high cost in data processing. This experiment was conducted to solve these revealing issues, by building a high performance yet cost-effective model for predicting accurate sentiments from large datasets containing customer reviews. This model uses the fastText library from Facebook’s AI research (FAIR) Lab, as well as the traditional Linear Support Vector Machine (LSVM) to classify text and word embedding. Comparisons of this model were also done with the author’s a custom multi-layer Sentiment Analysis (SA) Bi-directional Long Short-Term Memory (SA-BLSTM) model. The proposed fastText model, based on results, obtains a higher accuracy of 90.71% as well as 20% in performance compared to LSVM and SA-BLSTM models.

Human feelings are fundamental to perceive the conduct and state of mind of an individual. A healthy emotional state is one significant highlight to improve personal satisfaction. On the other hand, bad emotional health can prompt social or psychological well-being issues. Recognizing or detecting feelings in online health care data gives important and helpful information regarding the emotional state of patients. To recognize or detection of patient’s emotion against a specific disease using text from online sources is a challenging task. In this paper, we propose a method for the automatic detection of patient’s emotions in healthcare data using supervised machine learning approaches. For this purpose, we created a new dataset named EmoHD, comprising of 4,202 text samples against eight disease classes and six emotion classes, gathered from different online resources. We used six different supervised machine learning models based on different feature engineering techniques. We also performed a detailed comparison of the chosen six machine learning algorithms using different feature vectors on our dataset. We achieved the highest 87% accuracy using MultiLayer Perceptron as compared to other state of the art models. Moreover, we use the emotional guidance scale to show that there is a link between negative emotion and psychological health issues. Our proposed work will be helpful to automatically detect a patient’s emotion during disease and to avoid extreme acts like suicide, mental disorders, or psychological health issues. The implementation details are made publicly available at the given link: https://bit.ly/2NQeGET.

Speech emotion recognition (SER) is a challenging issue because it is not clear which features are effective for classification. Emotionally related features are always extracted from speech signals for emotional classification. Handcrafted features are mainly used for emotional identification from audio signals. However, these features are not sufficient to correctly identify the emotional state of the speaker. The advantages of a deep convolutional neural network (DCNN) are investigated in the proposed work. A pretrained framework is used to extract the features from speech emotion databases. In this work, we adopt the feature selection (FS) approach to find the discriminative and most important features for SER. Many algorithms are used for the emotion classification problem. We use the random forest (RF), decision tree (DT), support vector machine (SVM), multilayer perceptron classifier (MLP), and k-nearest neighbors (KNN) to classify seven emotions. All experiments are performed by utilizing four different publicly accessible databases. Our method obtains accuracies of 92.02%, 88.77%, 93.61%, and 77.23% for Emo-DB, SAVEE, RAVDESS, and IEMOCAP, respectively, for speaker-dependent (SD) recognition with the feature selection method. Furthermore, compared to current handcrafted feature-based SER methods, the proposed method shows the best results for speaker-independent SER. For EMO-DB, all classifiers attain an accuracy of more than 80% with or without the feature selection technique.

Text classification is a fundamental task in many applications such as topic labeling, sentiment analysis, and spam detection. The text syntactic relationship and word sequence are important and useful for text classification. How to model and incorporate them to improve performance is one key challenge. Inspired by human behavior in understanding text. In this paper, we combine the syntactic relationship, sequence structure, and semantics for text representation, and propose an attention-enhanced capsule network-based text classification model. Specifically, we use graph convolutional neural networks to encode syntactic dependency trees, build multi-head attention to encode dependencies relationship in text sequence, merge with semantic information by capsule network at last. Extensive experiments on five datasets demonstrate that our approach can effectively improve the performance of text classification compared with state-of-the-art methods. The result also shows capsule network, graph convolutional neural network, and multi-headed attention has integration effects on text classification tasks.

Different fields such as linguistics, teaching, and computing have demonstrated special interest in the study of sign languages (SL). However, the processes of teaching and learning these languages turn complex since it is unusual to find people teaching these languages that are fluent in both SL and the native language of the students. The teachings from deaf individuals become unique. Nonetheless, it is important for the student to lean on supportive mechanisms while being in the process of learning an SL. Bidirectional communication between deaf and hearing people through SL is a hot topic to achieve a higher level of inclusion. However, all the processes that convey teaching and learning SL turn difficult and complex since it is unusual to find SL teachers that are fluent also in the native language of the students, making it harder to provide computer teaching tools for different SL. Moreover, the main aspects that a second language learner of an SL finds difficult are phonology, non-manual components, and the use of space (the latter two are specific to SL, not to spoken languages). This proposal appears to be the first of the kind to favor the Costa Rican Sign Language (LESCO, for its Spanish acronym), as well as any other SL. Our research focus stands on reinforcing the learning process of final-user hearing people through a modular architectural design of a learning environment, relying on the concept of phonological proximity within a graphical tool with a high degree of usability. The aim of incorporating phonological proximity is to assist individuals in learning signs with similar handshapes. This architecture separates the logic and processing aspects from those associated with the access and generation of data, which makes it portable to other SL in the future. The methodology used consisted of defining 26 phonological parameters (13 for each hand), thus characterizing each sign appropriately. Then, a similarity formula was applied to compare each pair of signs. With these pre-calculations, the tool displays each sign and its top ten most similar signs. A SUS usability test and an open qualitative question were applied, as well as a numerical evaluation to a group of learners, to validate the proposal. In order to reach our research aims, we have analyzed previous work on proposals for teaching tools meant for the student to practice SL, as well as previous work on the importance of phonological proximity in this teaching process. This previous work justifies the necessity of our proposal, whose benefits have been proved through the experimentation conducted by different users on the usability and usefulness of the tool. To meet these needs, homonymous words (signs with the same starting handshape) and paronyms (signs with highly similar handshape), have been included to explore their impact on learning. It allows the possibility to apply the same perspective of our existing line of research to other SL in the future. This work was supported by the Spanish Ministry of Science, Innovation and Universities through the Project ECLIPSE-UA under Grant RTI2018-094283-B-C32, the Spanish Ministry of Science and Innovation through the Project AETHER-UA under Grant PID2020-112540RB-C43, the Project INTEGER under Grant RTI2018-094649-B-I00, and by the Conselleria de Educación, Investigación, Cultura y Deporte of the Community of Valencia, Spain, within the Project PROMETEO/2018/089, the School of Computing and the Computer Research Center at Costa Rica Institute of Technology and CONICIT (Consejo Nacional para Investigaciones Científicas y Tecnológicas), Costa Rica, under grant 290-2006.

Hand-written text recognition is useful for interpreting records in different fields such as healthcare, surgery and police in which professionals may avoid technical equipment and prefer writing notes on paper. In order to perform data fusion from different data sources, handwriting automatic recognition involves barriers such as different ways of writing letters and deformation due to many reasons. This work presents a novel handwriting recognition approach based on the application of coordinate vectors to find similarities in different kinds of deformations. In particular, it has been implemented using 16 segments in order to distinguish all the particularities in matching the new text considering a dataset with a machine-learning approach. The implementation of this approach with MATLAB shows promising results with accuracy of 92.8% for with ensemble and bagged trees, after analyzing 22 possible combinations of machine learning and processing techniques.

The COVID-19 pandemic is changing daily routines for many citizens with a high impact on the economy in some sectors. Small-medium enterprises of some sectors need to be aware of both the pandemic evolution and the corresponding sentiments of customers in order to figure out which are the best commercialization techniques. This article proposes an expert system based on the combination of machine learning and sentiment analysis in order to support business decisions with data fusion through web scraping. The system uses human-centric artificial intelligence for automatically generating explanations. The expert system feeds from online content from different sources using a scraping module. It allows users to interact with the expert system providing feedback, and the system uses this feedback to improve its recommendations with supervised learning.

Question answering (QA) is a hot field of research in Natural Language Processing. A big challenge in this field is to answer questions from knowledge-dependable domain. Since traditional QA hardly satisfies some knowledge-dependable situations, such as disease diagnosis, drug recommendation, etc. In recent years, researches focus on knowledge-based question answering (KBQA). However, there still exist some problems in KBQA, traditional KBQA is limited by a range of historical cases and takes too much human labor. To address the problems, in this paper, we propose an approach of knowledge graph based question answering (KGQA) method for medical domain, which firstly constructs a medical knowledge graph by extracting named entities and relations between the entities from medical documents. Then, in order to understand a question, it extracts the key information in the question according to the named entities, and meanwhile, it recognizes the questions’ intentions by adopting information gain. The next an inference method based on weighted path ranking on the knowledge graph is proposed to score the related entities according to the key information and intention of a given question. Finally, it extracts the inferred candidate entities to construct answers. Our approach can understand questions, connect the questions to the knowledge graph and inference the answers on the knowledge graph. Theoretical analysis and real-life experimental results show the efficiency of our approach.

DataStream mining is a challenging task for researchers because of the change in data distribution during classification, known as concept drift. Drift detection algorithms emphasize detecting the drift. The drift detection algorithm needs to be very sensitive to change in data distribution for detecting the maximum number of drifts in the data stream. But highly sensitive drift detectors lead to higher false-positive drift detections. This paper proposed a Drift Detection-based Adaptive Ensemble classifier for sentiment analysis and opinion mining, which uses these false-positive drift detections to benefit and minimize the negative impact of false-positive drift detection signals. The proposed method creates and adds a new classifier to the ensemble whenever a drift happens. A weighting mechanism is implemented, which provides weights to each classifier in the ensemble. The weight of the classifier decides the contribution of each classifier in the final classification results. The experiments are performed using different classification algorithms, and results are evaluated on the accuracy, precision, recall, and F1-measures. The proposed method is also compared with these state-of-the-art methods, OzaBaggingADWINClassifier, Accuracy Weighted Ensemble, Additive Expert Ensemble, Streaming Random Patches, and Adaptive Random Forest Classifier. The results show that the proposed method handles both true positive and false positive drifts efficiently.

Question classification is one of the essential tasks for automatic question answering implementation in natural language processing (NLP). Recently, there have been several text-mining issues such as text classification, document categorization, web mining, sentiment analysis, and spam filtering that have been successfully achieved by deep learning approaches. In this study, we illustrated and investigated our work on certain deep learning approaches for question classification tasks in an extremely inflected Turkish language. In this study, we trained and tested the deep learning architectures on the questions dataset in Turkish. In addition to this, we used three main deep learning approaches (Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNN)) and we also applied two different deep learning combinations of CNN-GRU and CNN-LSTM architectures. Furthermore, we applied the Word2vec technique with both skip-gram and CBOW methods for word embedding with various vector sizes on a large corpus composed of user questions. By comparing analysis, we conducted an experiment on deep learning architectures based on test and 10-cross fold validation accuracy. Experiment results were obtained to illustrate the effectiveness of various Word2vec techniques that have a considerable impact on the accuracy rate using different deep learning approaches. We attained an accuracy of 93.7% by using these techniques on the question dataset.

Sarcasm emerges as a common phenomenon across social networking sites because people express their negative thoughts, hatred and opinions using positive vocabulary which makes it a challenging task to detect sarcasm. Although various studies have investigated the sarcasm detection on baseline datasets, this work is the first to detect sarcasm from a multi-domain dataset that is constructed by combining Twitter and News Headlines datasets. This study proposes a hybrid approach where the convolutional neural networks (CNN) are used for feature extraction while the long short-term memory (LSTM) is trained and tested on those features. For performance analysis, several machine learning algorithms such as random forest, support vector classifier, extra tree classifier and decision tree are used. The performance of both the proposed model and machine learning algorithms is analyzed using the term frequency-inverse document frequency, bag of words approach, and global vectors for word representations. Experimental results indicate that the proposed model surpasses the performance of the traditional machine learning algorithms with an accuracy of 91.60%. Several state-of-the-art approaches for sarcasm detection are compared with the proposed model and results suggest that the proposed model outperforms these approaches concerning the precision, recall and F1 scores. The proposed model is accurate, robust, and performs sarcasm detection on a multi-domain dataset.

The success of supervised learning techniques for automatic speech processing does not always extend to problems with limited annotated speech. Unsupervised representation learning aims at utilizing unlabelled data to learn a transformation that makes speech easily distinguishable for classification tasks, whereby deep auto-encoder variants have been most successful in finding such representations. This paper proposes a novel mechanism to incorporate geometric position of speech samples within the global structure of an unlabelled feature set. Regression to the geometric position is also added as an additional constraint for the representation learning auto-encoder. The representation learnt by the proposed model has been evaluated over a supervised classification task for limited vocabulary keyword spotting, with the proposed representation outperforming the commonly used cepstral features by about 9% in terms of classification accuracy, despite using a limited amount of labels during supervision. Furthermore, a small keyword dataset has been collected for Kadazan, an indigenous, low-resourced Southeast Asian language. Analysis for the Kadazan dataset also confirms the superiority of the proposed representation for limited annotation. The results are significant as they confirm that the proposed method can learn unsupervised speech representations effectively for classification tasks with scarce labelled data.

Research of the techniques for effective fake news detection has become very needed and attractive. These techniques have a background in many research disciplines, including morphological analysis. Several researchers stated that simple content-related n-grams and POS tagging had been proven insufficient for fake news classification. However, they did not realise any empirical research results, which could confirm these statements experimentally in the last decade. Considering this contradiction, the main aim of the paper is to experimentally evaluate the potential of the common use of n-grams and POS tags for the correct classification of fake and true news. The dataset of published fake or real news about the current Covid-19 pandemic was pre-processed using morphological analysis. As a result, n-grams of POS tags were prepared and further analysed. Three techniques based on POS tags were proposed and applied to different groups of n-grams in the pre-processing phase of fake news detection. The n-gram size was examined as the first. Subsequently, the most suitable depth of the decision trees for sufficient generalization was scoped. Finally, the performance measures of models based on the proposed techniques were compared with the standardised reference TF-IDF technique. The performance measures of the model like accuracy, precision, recall and f1-score are considered, together with the 10-fold cross-validation technique. Simultaneously, the question, whether the TF-IDF technique can be improved using POS tags was researched in detail. The results showed that the newly proposed techniques are comparable with the traditional TF-IDF technique. At the same time, it can be stated that the morphological analysis can improve the baseline TF-IDF technique. As a result, the performance measures of the model, precision for fake news and recall for real news, were statistically significantly improved.

Navigation based task-oriented dialogue systems provide users with a natural way of communicating with maps and navigation software. Natural language understanding (NLU) is the first step for a task-oriented dialogue system. It extracts the important entities (slot tagging) from the user’s utterance and determines the user’s objective (intent determination). Word embeddings are the distributed representations of the input sentence, and encompass the sentence’s semantic and syntactic representations. We created the word embeddings using different methods like FastText, ELMO, BERT and XLNET; and studied their effect on the natural language understanding output. Experiments are performed on the Roman Urdu navigation utterances dataset. The results show that for the intent determination task XLNET based word embeddings outperform other methods; while for the task of slot tagging FastText and XLNET based word embeddings have much better accuracy in comparison to other approaches.

Recently, many users prefer online shopping to purchase items from the web. Shopping websites allow customers to submit comments and provide their feedback for the purchased products. Opinion mining and sentiment analysis are used to analyze products’ comments to help sellers and purchasers decide to buy products or not. However, the nature of online comments affects the performance of the opinion mining process because they may contain negation words or unrelated aspects to the product. To address these problems, a semantic-based aspect level opinion mining (SALOM) model is proposed. The SALOM extracts the product aspects based on the semantic similarity and classifies the comments. The proposed model considers the negation words and other types of product aspects such as aspects’ synonyms, hyponyms, and hypernyms to improve the accuracy of classification. Three different datasets are used to evaluate the proposed SALOM. The experimental results are promising in terms of Precision, Recall, and F-measure. The performance reaches 94.8% precision, 93% recall, and 92.6% f-measure.

Nowadays, social media experience an increase in hostility, which leads to many people suffering from online abusive behavior and harassment. We introduce a new publicly available annotated dataset for abusive language detection in short texts. The dataset includes comments from YouTube, along with contextual information: replies, video, video title, and the original description. The comments in the dataset are labeled as abusive or not and are classified by topic: politics, religion, and other. In particular, we discuss our refined annotation guidelines for such classification. We report a number of strong baselines on this dataset for the tasks of abusive language detection and topic classification, using a number of classifiers and text representations. We show that taking into account the conversational context, namely, replies, greatly improves the classification results as compared with using only linguistic features of the comments. We also study how the classification accuracy depends on the topic of the comment.

Baybayin is a pre-Hispanic Philippine writing system used in Luzon island. With the effort in reintroducing the script, in 2018, the Committee on Basic Education and Culture of the Philippine Congress approved House Bill 1022 or the ”National Writing System Act,” which declares the Baybayin script as the Philippines’ national writing system. Since then, Baybayin OCR has become a field of research interest. Numerous works have proposed different techniques in recognizing Baybayin scripts. However, all those studies anchored on the classification and recognition at the character level. In this work, we propose an algorithm that provides the Latin transliteration of a Baybayin word in an image. The proposed system relies on a Baybayin character classifier generated using the Support Vector Machine (SVM). The method involves isolation of each Baybayin character, then classifying each character according to its equivalent syllable in Latin script, and finally concatenate each result to form the transliterated word. The system was tested using a novel dataset of Baybayin word images and achieved a competitive 97.9% recognition accuracy. Based on our review of the literature, this is the first work that recognizes Baybayin scripts at the word level. The proposed system can be used in automated transliterations of Baybayin texts transcribed in old books, tattoos, signage, graphic designs, and documents, among others.

The evolution of electronic media is a mixed blessing. Due to the easy access, low cost, and faster reach of the information, people search out and devour news from online social networks. In contrast, the increasing acceptance of social media reporting leads to the spread of fake news. This is a minacious problem that causes disputes and endangers the societal stability and harmony. Fake news spread has gained attention from researchers due to its vicious nature. proliferation of misinformation in all media, from the internet to cable news, paid advertising and local news outlets, has made it essential for people to identify the misinformation and sort through the facts. Researchers are trying to analyze the credibility of information and curtail false information on such platforms. Credibility is the believability of the piece of information at hand. Analyzing the credibility of fake news is challenging due to the intent of its creation and the polychromatic nature of the news. In this work, we propose a model for detecting fake news. Our method investigates the content of the news at the early stage i.e., when the news is published but is yet to be disseminated through social media. Our work interprets the content with automatic feature extraction and the relevance of the text pieces. In summary, we introduce stance as one of the features along with the content of the article and employ the pre-trained contextualized word embeddings BERT to obtain the state-of-art results for fake news detection. The experiment conducted on the real-world dataset indicates that our model outperforms the previous work and enables fake news detection with an accuracy of 95.32%.

In the current age of overwhelming information and massive production of textual data on the Web, Event Detection has become an increasingly important task in various application domains. Several research branches have been developed to tackle the problem from different perspectives, including Natural Language Processing and Big Data analysis, with the goal of providing valuable resources to support decision-making in a wide variety of fields. In this paper, we propose a real-time domain-specific clustering-based event-detection approach that integrates textual information coming, on one hand, from traditional newswires and, on the other hand, from microblogging platforms. The goal of the implemented pipeline is twofold: (i) providing insights to the user about the relevant events that are reported in the press on a daily basis; (ii) alerting the user about potentially important and impactful events, referred to as hot events, for some specific tasks or domains of interest. The algorithm identifies clusters of related news stories published by globally renowned press sources, which guarantee authoritative, noise-free information about current affairs; subsequently, the content extracted from microblogs is associated to the clusters in order to gain an assessment of the relevance of the event in the public opinion. To identify the events of a day d we create the lexicon by looking at news articles and stock data of previous days up to d−1 Although the approach can be extended to a variety of domains (e.g. politics, economy, sports), we hereby present a specific implementation in the financial sector. We validated our solution through a qualitative and quantitative evaluation, performed on the Dow Jones’ Data, News and Analytics dataset, on a stream of messages extracted from the microblogging platform Stocktwits, and on the Standard & Poor’s 500 index time-series. The experiments demonstrate the effectiveness of our proposal in extracting meaningful information from real-world events and in spotting hot events in the financial sphere. An added value of the evaluation is given by the visual inspection of a selected number of significant real-world events, starting from the Brexit Referendum and reaching until the recent outbreak of the Covid-19 pandemic in early 2020.

Natural language inference (NLI) is an essential subtask in many natural language processing applications. It is a directional relationship from premise to hypothesis. A pair of texts is defined as entailed if a text infers its meaning from the other text. The NLI is also known as textual entailment recognition, and it recognizes entailed and contradictory sentences in various NLP systems like Question Answering, Summarization and Information retrieval systems. This paper describes the NLI problem attempted for a low resource Indian language Malayalam, the regional language of Kerala. More than 30 million people speak this language. The paper is about the Malayalam NLI dataset, named MaNLI dataset, and its application of NLI in Malayalam language using different models, namely Doc2Vec (paragraph vector), fastText, BERT (Bidirectional Encoder Representation from Transformers), and LASER (Language Agnostic Sentence Representation). Our work attempts NLI in two ways, as binary classification and as multiclass classification. For both the classifications, LASER outperformed the other techniques. For multiclass classification, NLI using LASER based sentence embedding technique outperformed the other techniques by a significant margin of 12% accuracy. There was also an accuracy improvement of 9% for LASER based NLI system for binary classification over the other techniques.

In an interactive online learning system (OLS), it is crucial for the learners to form the questions correctly in order to be provided or recommended appropriate learning materials. The incorrect question formation may lead the OLS to be confused, resulting in providing or recommending inappropriate study materials, which, in turn, affects the learning quality and experience and learner satisfaction. In this paper, we propose a novel method to assess the correctness of the learner's question in terms of syntax and semantics. Assessing the learner’s query precisely will improve the performance of the recommendation. A tri-gram language model is built, and trained and tested on corpora of 2,533 and 634 questions on Java, respectively, collected from books, blogs, websites, and university exam papers. The proposed method has exhibited 92% accuracy in identifying a question as correct or incorrect. Furthermore, in case the learner's input question is not correct, we propose an additional framework to guide the learner leading to a correct question that closely matches her intended question. For recommending correct questions, soft cosine based similarity is used. The proposed framework is tested on a group of learners' real-time questions and observed to accomplish 85% accuracy.

Emotion recognition in conversations is an important step in various virtual chat bots which require opinion-based feedback, like in social media threads, online support and many more applications. Current Emotion recognition in conversations models face issues like (a) loss of contextual information in between two dialogues of a conversation, (b) failure to give appropriate importance to significant tokens in each utterance and (c) inability to pass on the emotional information from previous utterances.The proposed model of Advanced Contextual Feature Extraction (AdCOFE) addresses these issues by performing unique feature extraction using knowledge graphs, sentiment lexicons and phrases of natural language at all levels (word and position embedding) of the utterances. Experiments on the Emotion recognition in conversations dataset show that AdCOFE is beneficial in capturing emotions in conversations., 12 pages, to be published in PeerJ Computer Science Journal

Hierarchical topic modeling is a potentially powerful instrument for determining topical structures of text collections that additionally allows constructing a hierarchy representing the levels of topic abstractness. However, parameter optimization in hierarchical models, which includes finding an appropriate number of topics at each level of hierarchy, remains a challenging task. In this paper, we propose an approach based on Renyi entropy as a partial solution to the above problem. First, we introduce a Renyi entropy-based metric of quality for hierarchical models. Second, we propose a practical approach to obtaining the “correct” number of topics in hierarchical topic models and show how model hyperparameters should be tuned for that purpose. We test this approach on the datasets with the known number of topics, as determined by the human mark-up, three of these datasets being in the English language and one in Russian. In the numerical experiments, we consider three different hierarchical models: hierarchical latent Dirichlet allocation model (hLDA), hierarchical Pachinko allocation model (hPAM), and hierarchical additive regularization of topic models (hARTM). We demonstrate that the hLDA model possesses a significant level of instability and, moreover, the derived numbers of topics are far from the true numbers for the labeled datasets. For the hPAM model, the Renyi entropy approach allows determining only one level of the data structure. For hARTM model, the proposed approach allows us to estimate the number of topics for two levels of hierarchy.

Multitask learning has led to significant advances in Natural Language Processing, including the decaNLP benchmark where question answering is used to frame 10 natural language understanding tasks in a single model. In this work we show how models trained to solve decaNLP fail with simple paraphrasing of the question. We contribute a crowd-sourced corpus of paraphrased questions (PQ-decaNLP), annotated with paraphrase phenomena. This enables analysis of how transformations such as swapping the class labels and changing the sentence modality lead to a large performance degradation. Training both MQAN and the newer T5 model using PQ-decaNLP improves their robustness and for some tasks improves the performance on the original questions, demonstrating the benefits of a model which is more robust to paraphrasing. Additionally, we explore how paraphrasing knowledge is transferred between tasks, with the aim of exploiting the multitask property to improve the robustness of the models. We explore the addition of paraphrase detection and paraphrase generation tasks, and find that while both models are able to learn these new tasks, knowledge about paraphrasing does not transfer to other decaNLP tasks.

In Information Retrieval (IR), Data Mining (DM), and Machine Learning (ML), similarity measures have been widely used for text clustering and classification. The similarity measure is the cornerstone upon which the performance of most DM and ML algorithms is completely dependent. Thus, till now, the endeavor in literature for an effective and efficient similarity measure is still immature. Some recently-proposed similarity measures were effective, but have a complex design and suffer from inefficiencies. This work, therefore, develops an effective and efficient similarity measure of a simplistic design for text-based applications. The measure developed in this work is driven by Boolean logic algebra basics (BLAB-SM), which aims at effectively reaching the desired accuracy at the fastest run time as compared to the recently developed state-of-the-art measures. Using the term frequency–inverse document frequency (TF-IDF) schema, the K-nearest neighbor (KNN), and the K-means clustering algorithm, a comprehensive evaluation is presented. The evaluation has been experimentally performed for BLAB-SM against seven similarity measures on two most-popular datasets, Reuters-21 and Web-KB. The experimental results illustrate that BLAB-SM is not only more efficient but also significantly more effective than state-of-the-art similarity measures on both classification and clustering tasks.

Video captioning,i.e., the task of generating captions from video sequences creates a bridge between the Natural Language Processing and Computer Vision domains of computer science. The task of generating a semantically accurate description of a video is quite complex. Considering the complexity, of the problem, the results obtained in recent research works are praiseworthy. However, there is plenty of scope for further investigation. This paper addresses this scope and proposes a novel solution. Most video captioning models comprise two sequential/recurrent layers—one as a video-to-context encoder and the other as a context-to-caption decoder. This paper proposes a novel architecture, namely Semantically Sensible Video Captioning (SSVC) which modifies the context generation mechanism by using two novel approaches—“stacked attention” and “spatial hard pull”. As there are no exclusive metrics for evaluating video captioning models, we emphasize both quantitative and qualitative analysis of our model. Hence, we have used the BLEU scoring metric for quantitative analysis and have proposed a human evaluation metric for qualitative analysis, namely the Semantic Sensibility (SS) scoring metric. SS Score overcomes the shortcomings of common automated scoring metrics. This paper reports that the use of the aforementioned novelties improves the performance of state-of-the-art architectures.

Science across all disciplines has become increasingly data-driven, leading to additional needs with respect to software for collecting, processing and analysing data. Thus, transparency about software used as part of the scientific process is crucial to understand provenance of individual research data and insights, is a prerequisite for reproducibility and can enable macro-analysis of the evolution of scientific methods over time. However, missing rigor in software citation practices renders the automated detection and disambiguation of software mentions a challenging problem. In this work, we provide a large-scale analysis of software usage and citation practices facilitated through an unprecedented knowledge graph of software mentions and affiliated metadata generated through supervised information extraction models trained on a unique gold standard corpus and applied to more than 3 million scientific articles. Our information extraction approach distinguishes different types of software and mentions, disambiguates mentions and outperforms the state-of-the-art significantly, leading to the most comprehensive corpus of 11.8 M software mentions that are described through a knowledge graph consisting of more than 300 M triples. Our analysis provides insights into the evolution of software usage and citation patterns across various fields, ranks of journals, and impact of publications. Whereas, to the best of our knowledge, this is the most comprehensive analysis of software use and citation at the time, all data and models are shared publicly to facilitate further research into scientific use and citation of software.

Background Fine-grained sentiment analysis is used to interpret consumers’ sentiments, from their written comments, towards specific entities on specific aspects. Previous researchers have introduced three main tasks in this field (ABSA, TABSA, MEABSA), covering all kinds of social media data (e.g., review specific, questions and answers, and community-based). In this paper, we identify and address two common challenges encountered in these three tasks, including the low-resource problem and the sentiment polarity bias. Methods We propose a unified model called PEA by integrating data augmentation methodology with the pre-trained language model, which is suitable for all the ABSA, TABSA and MEABSA tasks. Two data augmentation methods, which are entity replacement and dual noise injection, are introduced to solve both challenges at the same time. An ensemble method is also introduced to incorporate the results of the basic RNN-based and BERT-based models. Results PEA shows significant improvements on all three fine-grained sentiment analysis tasks when compared with state-of-the-art models. It also achieves comparable results with what the baseline models obtain while using only 20% of their training data, which demonstrates its extraordinary performance under extreme low-resource conditions.

The real-time availability of the Internet has engaged millions of users around the world. The usage of regional languages is being preferred for effective and ease of communication that is causing multilingual data on social networks and news channels. People share ideas, opinions, and events that are happening globally i.e., sports, inflation, protest, explosion, and sexual assault, etc. in regional (local) languages on social media. Extraction and classification of events from multilingual data have become bottlenecks because of resource lacking. In this research paper, we presented the event classification task for the Urdu language text existing on social media and the news channels by using machine learning classifiers. The dataset contains more than 0.1 million (102,962) labeled instances of twelve (12) different types of events. The title, its length, and the last four words of a sentence are used as features to classify the events. The Term Frequency-Inverse Document Frequency (tf-idf) showed the best results as a feature vector to evaluate the performance of the six popular machine learning classifiers. Random Forest (RF) and K-Nearest Neighbor (KNN) are among the classifiers that out-performed among other classifiers by achieving 98.00% and 99.00% accuracy, respectively. The novelty lies in the fact that the features aforementioned are not applied, up to the best of our knowledge, in the event extraction of the text written in the Urdu language.

Deep learning belongs to the field of artificial intelligence, where machines perform tasks that typically require some kind of human intelligence. Similar to the basic structure of a brain, a deep learning algorithm consists of an artificial neural network, which resembles the biological brain structure. Mimicking the learning process of humans with their senses, deep learning networks are fed with (sensory) data, like texts, images, videos or sounds. These networks outperform the state-of-the-art methods in different tasks and, because of this, the whole field saw an exponential growth during the last years. This growth resulted in way over 10,000 publications per year in the last years. For example, the search engine PubMed alone, which covers only a sub-set of all publications in the medical field, provides already over 11,000 results in Q3 2020 for the search term 'deep learning', and around 90% of these results are from the last three years. Consequently, a complete overview over the field of deep learning is already impossible to obtain and, in the near future, it will potentially become difficult to obtain an overview over a subfield. However, there are several review articles about deep learning, which are focused on specific scientific fields or applications, for example deep learning advances in computer vision or in specific tasks like object detection. With these surveys as a foundation, the aim of this contribution is to provide a first high-level, categorized meta-survey of selected reviews on deep learning across different scientific disciplines. The categories (computer vision, language processing, medical informatics and additional works) have been chosen according to the underlying data sources (image, language, medical, mixed). In addition, we review the common architectures, methods, pros, cons, evaluations, challenges and future directions for every sub-category., Comment: 83 pages, 22 figures, 9 tables, 100 references

Bengali is a low-resource language that lacks tools and resources for various natural language processing (NLP) tasks, such as sentiment analysis or profanity identification. In Bengali, only the translated versions of English sentiment lexicons are available. Moreover, no dictionary exists for detecting profanity in Bengali social media text. This study introduces a Bengali sentiment lexicon, BengSentiLex, and a Bengali swear lexicon, BengSwearLex. For creating BengSentiLex, a cross-lingual methodology is proposed that utilizes a machine translation system, a review corpus, two English sentiment lexicons, pointwise mutual information (PMI), and supervised machine learning (ML) classifiers in various stages. A semi-automatic methodology is presented to develop BengSwearLex that leverages an obscene corpus, word embedding, and part-of-speech (POS) taggers. The performance of BengSentiLex compared with the translated English lexicons in three evaluation datasets. BengSentiLex achieves 5%–50% improvement over the translated lexicons. For identifying profanity, BengSwearLex achieves documentlevel coverage of around 85% in an document-level in the evaluation dataset. The experimental results imply that BengSentiLex and BengSwearLex are effective resources for classifying sentiment and identifying profanity in Bengali social media content, respectively.

Melody and lyrics, reflecting two unique human cognitive abilities, are usually combined in music to convey emotions. Although psychologists and computer scientists have made considerable progress in revealing the association between musical structure and the perceived emotions of music, the features of lyrics are relatively less discussed. Using linguistic inquiry and word count (LIWC) technology to extract lyric features in 2,372 Chinese songs, this study investigated the effects of LIWC-based lyric features on the perceived arousal and valence of music. First, correlation analysis shows that, for example, the perceived arousal of music was positively correlated with the total number of lyric words and the mean number of words per sentence and was negatively correlated with the proportion of words related to the past and insight. The perceived valence of music was negatively correlated with the proportion of negative emotion words. Second, we used audio and lyric features as inputs to construct music emotion recognition (MER) models. The performance of random forest regressions reveals that, for the recognition models of perceived valence, adding lyric features can significantly improve the prediction effect of the model using audio features only; for the recognition models of perceived arousal, lyric features are almost useless. Finally, by calculating the feature importance to interpret the MER models, we observed that the audio features played a decisive role in the recognition models of both perceived arousal and perceived valence. Unlike the uselessness of the lyric features in the arousal recognition model, several lyric features, such as the usage frequency of words related to sadness, positive emotions, and tentativeness, played important roles in the valence recognition model.

Framing is a process of emphasizing a certain aspect of an issue over the others, nudging readers or listeners towards different positions on the issue even without making a biased argument. {Here, we propose FrameAxis, a method for characterizing documents by identifying the most relevant semantic axes ("microframes") that are overrepresented in the text using word embedding. Our unsupervised approach can be readily applied to large datasets because it does not require manual annotations. It can also provide nuanced insights by considering a rich set of semantic axes. FrameAxis is designed to quantitatively tease out two important dimensions of how microframes are used in the text. \textit{Microframe bias} captures how biased the text is on a certain microframe, and \textit{microframe intensity} shows how actively a certain microframe is used. Together, they offer a detailed characterization of the text. We demonstrate that microframes with the highest bias and intensity well align with sentiment, topic, and partisan spectrum by applying FrameAxis to multiple datasets from restaurant reviews to political news.} The existing domain knowledge can be incorporated into FrameAxis {by using custom microframes and by using FrameAxis as an iterative exploratory analysis instrument.} Additionally, we propose methods for explaining the results of FrameAxis at the level of individual words and documents. Our method may accelerate scalable and sophisticated computational analyses of framing across disciplines., Comment: 24 pages; published in PeerJ CS

The satisfaction of employees is very important for any organization to make sufficient progress in production and to achieve its goals. Organizations try to keep their employees satisfied by making their policies according to employees’ demands which help to create a good environment for the collective. For this reason, it is beneficial for organizations to perform staff satisfaction surveys to be analyzed, allowing them to gauge the levels of satisfaction among employees. Sentiment analysis is an approach that can assist in this regard as it categorizes sentiments of reviews into positive and negative results. In this study, we perform experiments for the world’s big six companies and classify their employees’ reviews based on their sentiments. For this, we proposed an approach using lexicon-based and machine learning based techniques. Firstly, we extracted the sentiments of employees from text reviews and labeled the dataset as positive and negative using TextBlob. Then we proposed a hybrid/voting model named Regression Vector-Stochastic Gradient Descent Classifier (RV-SGDC) for sentiment classification. RV-SGDC is a combination of logistic regression, support vector machines, and stochastic gradient descent. We combined these models under a majority voting criteria. We also used other machine learning models in the performance comparison of RV-SGDC. Further, three feature extraction techniques: term frequency-inverse document frequency (TF-IDF), bag of words, and global vectors are used to train learning models. We evaluated the performance of all models in terms of accuracy, precision, recall, and F1 score. The results revealed that RV-SGDC outperforms with a 0.97 accuracy score using the TF-IDF feature due to its hybrid architecture.

BackgroundRumor detection is a popular research topic in natural language processing and data mining. Since the outbreak of COVID-19, related rumors have been widely posted and spread on online social media, which have seriously affected people’s daily lives, national economy, social stability, etc. It is both theoretically and practically essential to detect and refute COVID-19 rumors fast and effectively. As COVID-19 was an emergent event that was outbreaking drastically, the related rumor instances were very scarce and distinct at its early stage. This makes the detection task a typical few-shot learning problem. However, traditional rumor detection techniques focused on detecting existed events with enough training instances, so that they fail to detect emergent events such as COVID-19. Therefore, developing a new few-shot rumor detection framework has become critical and emergent to prevent outbreaking rumors at early stages.MethodsThis article focuses on few-shot rumor detection, especially for detecting COVID-19 rumors from Sina Weibo with only a minimal number of labeled instances. We contribute a Sina Weibo COVID-19 rumor dataset for few-shot rumor detection and propose a few-shot learning-based multi-modality fusion model for few-shot rumor detection. A full microblog consists of the source post and corresponding comments, which are considered as two modalities and fused with the meta-learning methods.ResultsExperiments of few-shot rumor detection on the collected Weibo dataset and the PHEME public dataset have shown significant improvement and generality of the proposed model.

Small sample learning aims to learn information about object categories from a single or a few training samples. This learning style is crucial for deep learning methods based on large amounts of data. The deep learning method can solve small sample learning through the idea of meta-learning “how to learn by using previous experience.” Therefore, this paper takes image classification as the research object to study how meta-learning quickly learns from a small number of sample images. The main contents are as follows: After considering the distribution difference of data sets on the generalization performance of measurement learning and the advantages of optimizing the initial characterization method, this paper adds the model-independent meta-learning algorithm and designs a multi-scale meta-relational network. First, the idea of META-SGD is adopted, and the inner learning rate is taken as the learning vector and model parameter to learn together. Secondly, in the meta-training process, the model-independent meta-learning algorithm is used to find the optimal parameters of the model. The inner gradient iteration is canceled in the process of meta-validation and meta-test. The experimental results show that the multi-scale meta-relational network makes the learned measurement have stronger generalization ability, which further improves the classification accuracy on the benchmark set and avoids the need for fine-tuning of the model-independent meta-learning algorithm.

Platforms that feature user-generated content (social media, online forums, newspaper comment sections etc.) have to detect and filter offensive speech within large, fast-changing datasets. While many automatic methods have been proposed and achieve good accuracies, most of these focus on the English language, and are hard to apply directly to languages in which few labeled datasets exist. Recent work has therefore investigated the use ofcross-lingual transfer learningto solve this problem, training a model in a well-resourced language and transferring to a less-resourced target language; but performance has so far been significantly less impressive. In this paper, we investigate the reasons for this performance drop, via a systematic comparison of pre-trained models and intermediate training regimes on five different languages. We show that using a better pre-trained language model results in a large gain in overall performance and in zero-shot transfer, and that intermediate training on other languages is effective when little target-language data is available. We then use multiple analyses of classifier confidence and language model vocabulary to shed light on exactly where these gains come from and gain insight into the sources of the most typical mistakes.

Hate speech is one type of harmful online content which directly attacks or promotes hate towards a group or an individual member based on their actual or perceived aspects of identity, such as ethnicity, religion, and sexual orientation. With online hate speech on the rise, its automatic detection as a natural language processing task is gaining increasing interest. However, it is only recently that it has been shown that existing models generalise poorly to unseen data. This survey paper attempts to summarise how generalisable existing hate speech detection models are and the reasons why hate speech models struggle to generalise, sums up existing attempts at addressing the main obstacles, and then proposes directions of future research to improve generalisation in hate speech detection.

Semantic Text Similarity (STS) has several and important applications in the field of Natural Language Processing (NLP). The Aim of this study is to investigate the effect of stemming on text similarity for Arabic language at sentence level. Several Arabic light and heavy stemmers as well as lemmatization algorithms are used in this study, with a total of 10 algorithms. Standard training and testing data sets are used from SemEval-2017 international workshop for Task 1, Track 1 Arabic (ar–ar). Different features are selected to study the effect of stemming on text similarity based on different similarity measures. Traditional machine learning algorithms are used such as Support Vector Machines (SVM), Stochastic Gradient Descent (SGD) and Naïve Bayesian (NB). Compared to the original text, using the stemmed and lemmatized documents in experiments achieve enhanced Pearson correlation results. The best results attained when using Arabic light Stemmer (ARLSTem) and Farasa light stemmers, Farasa and Qalsadi Lemmatizers and Tashaphyne heavy stemmer. The best enhancement was about 7.34% in Pearson correlation. In general, stemming considerably improves the performance of sentence text similarly for Arabic language. However, some stemmers make results worse than those for original text; they are Khoja heavy stemmer and AlKhalil light stemmer.

Today, increased attention is drawn towards network representation learning, a technique that maps nodes of a network into vectors of a low-dimensional embedding space. A network embedding constructed this way aims to preserve nodes similarity and other specific network properties. Embedding vectors can later be used for downstream machine learning problems, such as node classification, link prediction and network visualization. Naturally, some networks have text information associated with them. For instance, in a citation network, each node is a scientific paper associated with its abstract or title; in a social network, all users may be viewed as nodes of a network and posts of each user as textual attributes. In this work, we explore how combining existing methods of text and network embeddings can increase accuracy for downstream tasks and propose modifications to popular architectures to better capture textual information in network embedding and fusion frameworks.

The cross-modal retrieval (CMR) has attracted much attention in the research community due to flexible and comprehensive retrieval. The core challenge in CMR is the heterogeneity gap, which is generated due to different statistical properties of multi-modal data. The most common solution to bridge the heterogeneity gap is representation learning, which generates a common sub-space. In this work, we propose a framework called “Improvement of Deep Cross-Modal Retrieval (IDCMR)”, which generates real-valued representation. The IDCMR preserves both intra-modal and inter-modal similarity. The intra-modal similarity is preserved by selecting an appropriate training model for text and image modality. The inter-modal similarity is preserved by reducing modality-invariance loss. The mean average precision (mAP) is used as a performance measure in the CMR system. Extensive experiments are performed, and results show that IDCMR outperforms over state-of-the-art methods by a margin 4% and 2% relatively with mAP in the text to image and image to text retrieval tasks on MSCOCO and Xmedia dataset respectively.