An Semantic Similarity Matching Method for Chinese Medical Question Text

For the purpose of capturing the semantic information accurately and clarifying the user's questioning intention, this paper proposes a novel, ensemble deep architecture BERT-MSBiLSTM-Attentions (BMA) which uses the Bidirectional Encoder Representations from Transformers (BERT), Multi-layer Siamese Bi-directional Long Short Term Memory (MSBiLSTM) and dual attention mechanism (Attentions) in order to solve the current question semantic similarity matching problem in medical automatic question answering system. In the preprocessing part, we first obtain token-level and sentence-level embedding vectors that contain rich semantic representations of complete sentences. The fusion of more accurate and adequate semantic features obtained through Siamese recurrent network and dual attention network can effectively eliminate the effect of poor matching results due to the presence of certain non-canonical texts or the diversity of their expression ambiguities. To evaluate our model, we splice the dataset of Ping An Healthkonnect disease QA transfer learning competition and "public AI star” challenge - COVID-19 similar sentence judgment competition. Experimental results with CC19 dataset show that BMA network achieves significant performance improvements compared to existing methods. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

UltrasonicG: Highly Robust Gesture Recognition on Ultrasonic Devices

In the current critical situation of novel coronavirus, the use of contactless gesture recognition method can reduce human contact and decrease the probability of virus transmission. In this context, ultrasound-based sensing has been widely concerned for its slow propagation speed, low sampling rate, and easy access to devices. However, limited by the complexity of gestural movements and insufficient training data, the accuracy and robustness of gesture recognition are low. To solve this problem, we propose UltrasonicG, a system for highly robust gesture recognition on ultrasonic devices. The system first converts a single audio signal into a Doppler shift and subsequently extracts the feature values using the Residual Neural Network (ResNet34) and uses Bi-directional Long Short-Term Memory (Bi-LSTM) for gesture recognition. The method effectively improves the accuracy of gesture recognition by combining the information of feature dimension with time dimension. To overcome the challenge of insufficient dataset, we use data extension to expand the dataset. We have conducted extensive experiments and evaluations on UltrasonicG in a variety of real scenarios. The experimental results show that UltrasonicG can recognize 15 kinds of gestures with a recognition distance of 0.5 m. And it has a high accuracy and robustness with a comprehensive recognition rate of 98.8% under different environments and influencing factors. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Analysis of public sentiment tendency in COVID-19 period based on GA-BiLSTM

As the Internet becomes the main source of information for the public, grasping the emotional polarity of online public opinion is particularly important for relevant departments to supervise online public opinion. In order to more accurately determine the emotional polarity of public opinion in the epidemic, this paper proposes a public sentiment analysis model based on Word2vec, genetic algorithm and Bi-directional Long Short-Term Memory (Bi-LSTM) algorithm. The Word2vec model converts the comment text into an n-dimensional vector, uses the Bi-LSTM algorithm to analyze the sentiment polarity, and uses the genetic algorithm to analyze the number of Bi-LSTM layers and the number of fully connected layers and the number of neurons in each layer of Bi-LSTM optimization. The experimental results show that the accuracy of the above model is compared with the accuracy of the Word2vec model and the LSTM model separately, and the accuracy is increased by 11.0% and 7.7%, respectively. © 2022 ACM.

Predicting Organization Performance Changes: A Sequential Data-Based Framework.

The business environment is increasingly uncertain due to the rapid development of disruptive information technologies, the changing global economy, and the COVID-19 pandemic. This brings great uncertainties to investors to predict the performance changes and risks of companies. This research proposes a sequential data-based framework that aggregates data from multiple sources including both structured and unstructured data to predict the performance changes. It leverages data generated from the early risk warning system in China stock market to measure and predict organization performance changes based on the risk warning status changes of public companies. Different from the models in existing literature that focus on the prediction of risk warning of companies, our framework predicts a portfolio of organization performance changes, including business decline and recovery, thus helping investors to not only predict public company risks, but also discover investment opportunities. By incorporating sequential data, our framework achieves 92.3% macro-F1 value on real-world data from listed companies in China, outperforming other static models.

Deep Learning-Based Sentiment Analysis on COVID-19 News Videos

Coronavirus disease (COVID-19) has adversely affected all walks of human life. The whole world is confronting this deadly virus, and no country in this world remains untouched during this pandemic. There are several online news videos related to COVID-19 that are shared on various online platforms such as YouTube, DailyMotion, and Vimeo. There were several arguments on the genuineness of the contents, people watch them, share them, and most importantly express their views and opinions as comments on those platforms. Analyzing these comments can unearth the patterns hidden in them to study people's responses to videos on COVID-19. This paper proposes a deep learning-based sentiment analysis approach to people's response toward online COVID-19 video news. This work implements different deep learning approaches such as LSTM, Bi-LSTM, CNN, and GRU to classify sentiment from the comments collected from YouTube. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Bi-directional Long Short-Term Memory Network for Fake News Detection from Social Media

These days’ web-based media is one of the main news hotspots for individuals throughout the planet for its minimal expense, simple openness, and quick spreading. This web-based media can in some cases include uncertain messages and has a critical danger of openness to counterfeit or fake news, which may elude the pursuers. Therefore, finding fake news from social media is one of the important natural language processing tasks. In this work, we have proposed a bi-directional long short-term memory (Bi-LSTM) network to identify COVID-19 fake news posted on Twitter. The performance of the proposed Bi-LSTM network is compared to six different popular classical machine learning classifiers such as Naïve Bayes, KNN, Decision Tree, Gradient Boosting, Random Forest, and AdaBoost. In the case of classical machine learning classifiers uni-gram, bi-gram, and tri-gram word TF-IDF features are used whereas in the case of the Bi-LSTM model word embedding features are used. The proposed Bi-LSTM network performed best in comparison to other implemented models and achieved a weighted F1-score of 0.94 in identifying COVID-19 fake news from Twitter. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.