Multitask Learning Neural Networks for Pandemic Prediction with Public Stance Enhancement

State and local governments have imposed health policies to contain the spread of COVID-19 since it had a serious impact on human daily life. However, the public stance on these measures may be time-varying. It is likely to escalate the infection in the area where the public is negative or resistant. To take advantage of the correlation between public stance on health policies and the COVID-19 statistics, we propose a novel framework, Multitask Learning Neural Networks for Pandemic Prediction with Public Stance Enhancement (MP3), which is composed of three modules: (1) Stance awareness module to make stance detection on health policies from users' tweets in social media and convert them into a stance time series. (2) Temporal feature extraction module that applies Convolution Neural Network and Recurrent Neural Network to extract and fuse local patterns and long-term correlations from COVID-19 statistics. Moreover, a Stance Latency-aware Attention is proposed to capture dynamic social effects and fuse them with temporal features. (3) Multi-task prediction module to adopt Graph Convolution Network to model the spread of pandemic and employ multi-task learning to simultaneously predict COVID-19 statistics and the trend of public stance on health policies. The proposed framework outperforms state-of-the-art baselines on both confirmed cases and deaths prediction tasks. © 2022 IEEE.

Fine-tuning BERT to Classify COVID19 Tweets Containing Symptoms

Twitter provides a source of patient-generated data that has been used in various population health studies. The first step in many of these studies is to identify and capture Twitter messages (tweets) containing medication mentions. Identifying personal mentions of COVID19 symptoms requires distinguishing personal mentions from other mentions such as symptoms reported by others and references to news articles or other sources. In this article, we describe our submission to Task 6 of the Social Media Mining for Health Applications (SMM4H) Shared Task 2021. This task challenged participants to classify tweets where the target classes are - (1) self-reports, (2) non-personal reports, and (3) literature/news mentions. Our system uses a handcrafted preprocessing and word embeddings from BERT encoder model. We achieve. F1 score of 93%. © 2021 Association for Computational Linguistics.

ULD-NUIG at Social Media Mining for Health Applications (#SMM4H) Shared Task 2021

In this paper, we present the ULD-NUIG team's system, designed as part of Social Media Mining for Health Applications (#SMM4H) Shared Task 2021. We participate in two tasks out of eight, namely "Classification of tweets self-reporting potential cases of COVID-19" (Task 5) and "Classification of COVID19 tweets containing symptoms" (Task 6). The team conduct a series of experiments to explore the challenges of both the tasks. We used a multilingual pre-trained BERT model for Task 5 and Generative Morphemes with Attention (GenMA) model for Task 6. In the experiments, we find that, GenMA, developed for Task 6, gives better results on both validation and test data-set. The submitted systems achieve F-1 score 0.53 for Task 5 and 0.84 for Task 6 on test data-set. © 2021 Association for Computational Linguistics.

NLP@NISER: Classification of COVID19 tweets containing symptoms

In this paper, we describe our approaches for task six of Social Media Mining for Health Applications (SMM4H) shared task in 2021. The task is to classify twitter tweets containing COVID-19 symptoms in three classes (self-reports, non-personal reports & literature/news mentions). We implemented BERT and XLNet for this text classification task. Best result was achieved by XLNet approach, which is F1 score 0.94, precision 0.9448 and recall 0.94448. This is slightly better than the average score, i.e. F1 score 0.93, precision 0.93235 and recall 0.93235. © 2021 Association for Computational Linguistics.

Fine-Tuning Transformers for Identifying Self-Reporting Potential Cases and Symptoms of COVID-19 in Tweets

We describe our straight-forward approach for Tasks 5 and 6 of 2021 Social Media Mining for Health Applications (SMM4H) shared tasks. Our system is based on fine-tuning DistillBERT on each task, as well as first fine-tuning the model on the other task. We explore how much fine-tuning is necessary for accurately classifying tweets as containing self-reported COVID-19 symptoms (Task 5) or whether a tweet related to COVID-19 is self-reporting, non-personal reporting, or a literature/news mention of the virus (Task 6). © 2021 Association for Computational Linguistics.

Approaching SMM4H with auto-regressive language models and back-translation

We describe our submissions to the 6th edition of the Social Media Mining for Health Applications (SMM4H) shared task. Our team (OGNLP) participated in the sub-task: Classification of tweets self-reporting potential cases of COVID-19 (Task 5). For our submissions, we employed systems based on autoregressive transformer models (XLNet) and back-translation for balancing the dataset. © 2021 Association for Computational Linguistics.

Improving Text Clustering Using a New Technique for Selecting Trustworthy Content in Social Networks

Today’s information society has led to the emergence of a large number of applications that generate and consume digital data. Many of these applications are based on social networks, and therefore their information often comes in the form of unstructured text. This text from social media also tends to contain a high level of noise and untrustworthy content. Therefore, having systems capable of dealing with it efficiently is a very relevant issue. In order to verify the trustworthiness of the social media content, it is necessary to analyse and explore social media data by using text mining techniques. One of the most widespread techniques in the field of text mining is text clustering, that allows us to automatically group similar documents into categories. Text clustering is very sensitive to the presence of noise and so in this paper we propose a pre-processing pipeline based on word embedding that allows selecting trustworthy content and discarding noise in a way that improves clustering results. To validate the proposed pipeline, a real use case is provided on a Twitter dataset related to COVID-19. © 2022, Springer Nature Switzerland AG.