AI Blitz XIII Faces challenge hosted on www.aicrowd.com platform consisted of five problems: Sentiment Classification, Age Prediction, Mask Prediction, Face Recognition, and Face De-Blurring. Our team GLaDOS took second place. Here we present our solutions and results. Code implementation: https://github.com/ndrwmlnk/ai-blitz-xiii
With the constantly growing number of reviews and other sentiment-bearing texts on the Web, the demand for automatic sentiment analysis algorithms continues to expand. Aspect-based sentiment classification (ABSC) allows for the automatic extraction of highly fine-grained sentiment information from text documents or sentences. In this survey, the rapidly evolving state of the research on ABSC is reviewed. A novel taxonomy is proposed that categorizes the ABSC models into three major categories: knowledge-based, machine learning, and hybrid models. This taxonomy is accompanied with summarizing overviews of the reported model performances, and both technical and intuitive explanations of the various ABSC models. State-of-the-art ABSC models are discussed, such as models based on the transformer model, and hybrid deep learning models that incorporate knowledge bases. Additionally, various techniques for representing the model inputs and evaluating the model outputs are reviewed. Furthermore, trends in the research on ABSC are identified and a discussion is provided on the ways in which the field of ABSC can be advanced in the future.
As Natural Language Processing (NLP) algorithms continually achieve new milestones, out-of-distribution generalization remains a significant challenge. This paper addresses the issue of multi-source adaptation for unfamiliar domains: We leverage labeled data from multiple source domains to generalize to unknown target domains at training. Our innovative framework employs example-based Hypernetwork adaptation: a T5 encoder-decoder initially generates a unique signature from an input example, embedding it within the source domains' semantic space. This signature is subsequently utilized by a Hypernetwork to generate the task classifier's weights. We evaluated our method across two tasks - sentiment classification and natural language inference - in 29 adaptation scenarios, where it outpaced established algorithms. In an advanced version, the signature also enriches the input example's representation. We also compare our finetuned architecture to few-shot GPT-3, demonstrating its effectiveness in essential use cases. To our knowledge, this marks the first application of Hypernetworks to the adaptation for unknown domains.
Because of the current COVID-19 pandemic with its increasing fears among people, it has triggered several health complications such as depression and anxiety. Such complications have not only affected the developed countries but also developing countries such as Nepal. These complications can be understood from peoples' tweets/comments posted online after their proper analysis and sentiment classification. Nevertheless, owing to the limited number of tokens/words in each tweet, it is always crucial to capture multiple information associated with them for their better understanding. In this study, we, first, represent each tweet by combining both syntactic and semantic information, called hybrid features. The syntactic information is generated from the bag of words method, whereas the semantic information is generated from the combination of the fastText-based (ft) and domain-specific (ds) methods. Second, we design a novel multi-channel convolutional neural network (MCNN), which ensembles the multiple CNNs, to capture multi-scale information for better classification. Last, we evaluate the efficacy of both the proposed feature extraction method and the MCNN model classifying tweets into three sentiment classes (positive, neutral and negative) on NepCOV19Tweets dataset, which is the only public COVID-19 tweets dataset in Nepali language. The evaluation results show that the proposed hybrid features outperform individual feature extraction methods with the highest classification accuracy of 69.7% and the MCNN model outperforms the existing methods with the highest classification accuracy of 71.3% during classification.
Simile interpretation is a crucial task in natural language processing. Nowadays, pre-trained language models (PLMs) have achieved state-of-the-art performance on many tasks. However, it remains under-explored whether PLMs can interpret similes or not. In this paper, we investigate the ability of PLMs in simile interpretation by designing a novel task named Simile Property Probing, i.e., to let the PLMs infer the shared properties of similes. We construct our simile property probing datasets from both general textual corpora and human-designed questions, containing 1,633 examples covering seven main categories. Our empirical study based on the constructed datasets shows that PLMs can infer similes' shared properties while still underperforming humans. To bridge the gap with human performance, we additionally design a knowledge-enhanced training objective by incorporating the simile knowledge into PLMs via knowledge embedding methods. Our method results in a gain of 8.58% in the probing task and 1.37% in the downstream task of sentiment classification. The datasets and code are publicly available at https://github.com/Abbey4799/PLMs-Interpret-Simile.
The task of joint dialog sentiment classification (DSC) and act recognition (DAR) aims to simultaneously predict the sentiment label and act label for each utterance in a dialog. In this paper, we put forward a new framework which models the explicit dependencies via integrating \textit{prediction-level interactions} other than semantics-level interactions, more consistent with human intuition. Besides, we propose a speaker-aware temporal graph (SATG) and a dual-task relational temporal graph (DRTG) to introduce \textit{temporal relations} into dialog understanding and dual-task reasoning. To implement our framework, we propose a novel model dubbed DARER, which first generates the context-, speaker- and temporal-sensitive utterance representations via modeling SATG, then conducts recurrent dual-task relational reasoning on DRTG, in which process the estimated label distributions act as key clues in prediction-level interactions. Experiment results show that DARER outperforms existing models by large margins while requiring much less computation resource and costing less training time. Remarkably, on DSC task in Mastodon, DARER gains a relative improvement of about 25% over previous best model in terms of F1, with less than 50% parameters and about only 60% required GPU memory.
In this paper, we study sentiment analysis task where the outcomes are mainly contributed by a few key elements of the inputs. Motivated by the two-streams hypothesis, we propose a neural architecture, named TraceNet, to address this type of task. It not only learns discriminative representations for the target task via its encoders, but also traces key elements at the same time via its locators. In TraceNet, both encoders and locators are organized in a layer-wise manner, and a smoothness regularization is employed between adjacent encoder-locator combinations. Moreover, a sparsity constraints are enforced on locators for tracing purposes and items are proactively masked according to the item weights output by locators.A major advantage of TraceNet is that the outcomes are easier to understand, since the most responsible parts of inputs are identified. Also, under the guidance of locators, it is more robust to attacks due to its focus on key elements and the proactive masking training strategy. Experimental results show its effectiveness for sentiment classification. Moreover, we provide several case studies to demonstrate its robustness and interpretability.
Aspect-based sentiment classification (ABSC) is a very challenging subtask of sentiment analysis (SA) and suffers badly from the class-imbalance. Existing methods only process sentences independently, without considering the domain-level relationship between sentences, and fail to provide effective solutions to the problem of class-imbalance. From an intuitive point of view, sentences in the same domain often have high-level semantic connections. The interaction of their high-level semantic features can force the model to produce better semantic representations, and find the similarities and nuances between sentences better. Driven by this idea, we propose a plug-and-play Pairwise Semantic Interaction (PSI) module, which takes pairwise sentences as input, and obtains interactive information by learning the semantic vectors of the two sentences. Subsequently, different gates are generated to effectively highlight the key semantic features of each sentence. Finally, the adversarial interaction between the vectors is used to make the semantic representation of two sentences more distinguishable. Experimental results on four ABSC datasets show that, in most cases, PSI is superior to many competitive state-of-the-art baselines and can significantly alleviate the problem of class-imbalance.
Attacks on deep learning models are often difficult to identify and therefore are difficult to protect against. This problem is exacerbated by the use of public datasets that typically are not manually inspected before use. In this paper, we offer a solution to this vulnerability by using, during testing, random perturbations such as spelling correction if necessary, substitution by random synonym, or simply dropping the word. These perturbations are applied to random words in random sentences to defend NLP models against adversarial attacks. Our Random Perturbations Defense and Increased Randomness Defense methods are successful in returning attacked models to similar accuracy of models before attacks. The original accuracy of the model used in this work is 80% for sentiment classification. After undergoing attacks, the accuracy drops to accuracy between 0% and 44%. After applying our defense methods, the accuracy of the model is returned to the original accuracy within statistical significance.
Aspect-based sentiment analysis (ABSA) typically requires in-domain annotated data for supervised training/fine-tuning. It is a big challenge to scale ABSA to a large number of new domains. This paper aims to train a unified model that can perform zero-shot ABSA without using any annotated data for a new domain. We propose a method called contrastive post-training on review Natural Language Inference (CORN). Later ABSA tasks can be cast into NLI for zero-shot transfer. We evaluate CORN on ABSA tasks, ranging from aspect extraction (AE), aspect sentiment classification (ASC), to end-to-end aspect-based sentiment analysis (E2E ABSA), which show ABSA can be conducted without any human annotated ABSA data.
After transformer is proposed, lots of pre-trained language models have been come up with and sentiment analysis (SA) task has been improved. In this paper, we proposed a method that uses an auxiliary sentence about aspects that the sentence contains to help sentiment prediction. The first is aspect detection, which uses a multi-aspects detection model to predict all aspects that the sentence has. Combining the predicted aspects and the original sentence as Sentiment Analysis (SA) model's input. The second is to do out-of-domain aspect-based sentiment analysis(ABSA), train sentiment classification model with one kind of dataset and validate it with another kind of dataset. Finally, we created two baselines, they use no aspect and all aspects as sentiment classification model's input, respectively. Compare two baselines performance to our method, found that our method really makes sense.
Explaining machine learning models is an important and increasingly popular area of research interest. The Shapley value from game theory has been proposed as a prime approach to compute feature importance towards model predictions on images, text, tabular data, and recently graph neural networks (GNNs) on graphs. In this work, we revisit the appropriateness of the Shapley value for GNN explanation, where the task is to identify the most important subgraph and constituent nodes for GNN predictions. We claim that the Shapley value is a non-ideal choice for graph data because it is by definition not structure-aware. We propose a Graph Structure-aware eXplanation (GStarX) method to leverage the critical graph structure information to improve the explanation. Specifically, we define a scoring function based on a new structure-aware value from the cooperative game theory proposed by Hamiache and Navarro (HN). When used to score node importance, the HN value utilizes graph structures to attribute cooperation surplus between neighbor nodes, resembling message passing in GNNs, so that node importance scores reflect not only the node feature importance, but also the node structural roles. We demonstrate that GStarX produces qualitatively more intuitive explanations, and quantitatively improves explanation fidelity over strong baselines on chemical graph property prediction and text graph sentiment classification.
We propose a novel multi-task pre-training method for Speech Emotion Recognition (SER). We pre-train SER model simultaneously on Automatic Speech Recognition (ASR) and sentiment classification tasks to make the acoustic ASR model more ``emotion aware''. We generate targets for the sentiment classification using text-to-sentiment model trained on publicly available data. Finally, we fine-tune the acoustic ASR on emotion annotated speech data. We evaluated the proposed approach on the MSP-Podcast dataset, where we achieved the best reported concordance correlation coefficient (CCC) of 0.41 for valence prediction.
The smartphone apps market is growing rapidly which challenges apps owners to continue improving their products and to compete in the market. The analysis of users feedback is a key enabler for improvements as stakeholders can utilize it to gain a broad understanding of the successes and failures of their products as well as those of competitors. That leads to generating evidence-based requirements and enhancing the requirements elicitation activities. Aspect-Based Sentiment Analysis (ABSA) is a branch of Sentiment Analysis that identifies aspects and assigns a sentiment to each aspect. Having the aspect information adds a more accurate understanding of opinions and addresses the limited use of the overall sentiment. However, the ABSA task has not yet been investigated in the context of smartphone apps reviews and requirements elicitation. In this paper, we introduce AWARE as a benchmark dataset of 11323 apps reviews that are annotated with aspect terms, categories, and sentiment. Reviews were collected from three domains: productivity, social networking, and games. We derived the aspect categories for each domain using content analysis and validated them with domain experts in terms of importance, comprehensiveness, overlapping, and granularity level. We crowdsourced the annotations of aspect categories and sentiment polarities and performed quality control procedures. The aspect terms were annotated using a partially automated Natural Language Processing (NLP) approach and validated by annotators, which resulted in 98% correct aspect terms. Lastly, we built machine learning baselines for three tasks, namely (i) aspect term extraction using a POS tagger, (ii) aspect category classification, and (iii) aspect sentiment classification, using both Support Vector Machine (SVM) and Multi-layer Perceptron (MLP) classifiers.
Due to the rapid development of Internet-based applications such as social media, sentiment analysis has become one of the most widely used research areas of natural language processing and an important tool for extracting opinions from texts. Because a huge number of comments and reviews are generated today through social media, and these comments are very significant. However, manually analyzing and summarizing them requires a lot of time and money. Therefore, sentiment analysis has entered the field to organize and analyze them by creating an automated system. In recent years, the use of deep learning in sentiment analysis has shown powerful results. However, creating a model alone may not provide the best predictions and lead to errors such as bias and high variance. To reduce these errors and improve the efficiency of model predictions, combining several models known as ensemble learning may provide better results. Therefore, the main purpose of this article is to create a model based on ensemble learning using Several convolutional neural networks and majority voting classifiers to enhance sentiment analysis in Persian reviews. The proposed model was evaluated on two datasets of reviews of electronic products and movies by 5-fold and 10-fold cross-validation. The results indicate that this new approach increases the efficiency of the sentiment analysis model in Persian language.
In aspect-level sentiment classification (ASC), state-of-the-art models encode either syntax graph or relation graph to capture the local syntactic information or global relational information. Despite the advantages of syntax and relation graphs, they have respective shortages which are neglected, limiting the representation power in the graph modeling process. To resolve their limitations, we design a novel local-global interactive graph, which marries their advantages by stitching the two graphs via interactive edges. To model this local-global interactive graph, we propose a novel neural network termed DigNet, whose core module is the stacked local-global interactive (LGI) layers performing two processes: intra-graph message passing and cross-graph message passing. In this way, the local syntactic and global relational information can be reconciled as a whole in understanding the aspect-level sentiment. Concretely, we design two variants of local-global interactive graphs with different kinds of interactive edges and three variants of LGI layers. We conduct experiments on several public benchmark datasets and the results show that we outperform previous best scores by 3\%, 2.32\%, and 6.33\% in terms of Macro-F1 on Lap14, Res14, and Res15 datasets, respectively, confirming the effectiveness and superiority of the proposed local-global interactive graph and DigNet.
Domain adaptation tasks such as cross-domain sentiment classification aim to utilize existing labeled data in the source domain and unlabeled or few labeled data in the target domain to improve the performance in the target domain via reducing the shift between the data distributions. Existing cross-domain sentiment classification methods need to distinguish pivots, i.e., the domain-shared sentiment words, and non-pivots, i.e., the domain-specific sentiment words, for excellent adaptation performance. In this paper, we first design a Category Attention Network (CAN), and then propose a model named CAN-CNN to integrate CAN and a Convolutional Neural Network (CNN). On the one hand, the model regards pivots and non-pivots as unified category attribute words and can automatically capture them to improve the domain adaptation performance; on the other hand, the model makes an attempt at interpretability to learn the transferred category attribute words. Specifically, the optimization objective of our model has three different components: 1) the supervised classification loss; 2) the distributions loss of category feature weights; 3) the domain invariance loss. Finally, the proposed model is evaluated on three public sentiment analysis datasets and the results demonstrate that CAN-CNN can outperform other various baseline methods.
This paper studies continual learning (CL) for sentiment classification (SC). In this setting, the CL system learns a sequence of SC tasks incrementally in a neural network, where each task builds a classifier to classify the sentiment of reviews of a particular product category or domain. Two natural questions are: Can the system transfer the knowledge learned in the past from the previous tasks to the new task to help it learn a better model for the new task? And, can old models for previous tasks be improved in the process as well? This paper proposes a novel technique called KAN to achieve these objectives. KAN can markedly improve the SC accuracy of both the new task and the old tasks via forward and backward knowledge transfer. The effectiveness of KAN is demonstrated through extensive experiments.
Sentiment Analysis is one of the most classical and primarily studied natural language processing tasks. This problem had a notable advance with the proposition of more complex and scalable machine learning models. Despite this progress, the Brazilian Portuguese language still disposes only of limited linguistic resources, such as datasets dedicated to sentiment classification, especially when considering the existence of predefined partitions in training, testing, and validation sets that would allow a more fair comparison of different algorithm alternatives. Motivated by these issues, this work analyzes the predictive performance of a range of document embedding strategies, assuming the polarity as the system outcome. This analysis includes five sentiment analysis datasets in Brazilian Portuguese, unified in a single dataset, and a reference partitioning in training, testing, and validation sets, both made publicly available through a digital repository. A cross-evaluation of dataset-specific models over different contexts is conducted to evaluate their generalization capabilities and the feasibility of adopting a unique model for addressing all scenarios.
Background: The COVID-19 pandemic has affected our society and human well-being in various ways. In this study, we investigate how the pandemic has influenced people's emotions and psychological states compared to a pre-pandemic period using real-world data from social media. Method: We collected Reddit social media data from 2019 (pre-pandemic) and 2020 (pandemic) from the subreddits communities associated with eight universities. We applied the pre-trained Robustly Optimized BERT pre-training approach (RoBERTa) to learn text embedding from the Reddit messages, and leveraged the relational information among posted messages to train a graph attention network (GAT) for sentiment classification. Finally, we applied model stacking to combine the prediction probabilities from RoBERTa and GAT to yield the final classification on sentiment. With the model-predicted sentiment labels on the collected data, we used a generalized linear mixed-effects model to estimate the effects of pandemic and in-person teaching during the pandemic on sentiment. Results: The results suggest that the odds of negative sentiments in 2020 (pandemic) were 25.7% higher than the odds in 2019 (pre-pandemic) with a $p$-value $<0.001$; and the odds of negative sentiments associated in-person learning were 48.3% higher than with remote learning in 2020 with a $p$-value of 0.029. Conclusions: Our study results are consistent with the findings in the literature on the negative impacts of the pandemic on people's emotions and psychological states. Our study contributes to the growing real-world evidence on the various negative impacts of the pandemic on our society; it also provides a good example of using both ML techniques and statistical modeling and inference to make better use of real-world data.