Template for Extended Abstract contributions to INDOOR AIR 2022

This investigation presents results of Computational Fluid Dynamics (CFD) modelling of aerosol behaviour within an arbitrary 'realistic' 100m2 office environment, with dynamic and variable respiratory droplet release profile applied based on published findings (Morawska et al., 2009). A multitude of ventilation strategies and configurations have been applied to the base model to compare the effectiveness of reducing the concentration of suspended aerosols over time. A key finding of the investigation indicates a relatively low sensitivity to increasing outside air percentage, and that the benefit from this strategy is heavily dependent on the in-duct droplet decay factor. The application of local recirculating air filtration systems with MERV-13 filters mounted on occupant desks proved significantly more effectiveness than increasing outside air concentration from 25% to 100% in reducing the quantity of suspended aerosols. This highlights that the ventilation industry should perhaps focus on opportunities to integrate filtration systems into furniture, partitions, cabinetry etc., and that an appliance-based solution may be more beneficial for reducing COVID-19 transmission in buildings (and likely more straightforward) than modifications to central ventilation systems, particularly in the application of refurbishments and retrofits. © 2022 17th International Conference on Indoor Air Quality and Climate, INDOOR AIR 2022. All rights reserved.

COVID-QA: A question answering dataset for COVID-19

We present COVID-QA, a Question Answering dataset consisting of 2,019 question/answer pairs annotated by volunteer biomedical experts on scientific articles related to COVID-19. To evaluate the dataset we compared a RoBERTa base model fine-tuned on SQuAD with the same model trained on SQuAD and our COVID-QA dataset. We found that the additional training on this domain-specific data leads to significant gains in performance. Both the trained model and the annotated dataset have been open-sourced at: https://github.com/deepset-ai/COVID-QA. © ACL 2020.All right reserved.

Fusion of Periocular Deep Features in a Dual-Input CNN for Biometric Recognition

Periocular recognition has attracted attention in recent times. The advent of the COVID-19 pandemic and the consequent obligation to wear facial masks made face recognition problematic due to the important occlusion of the lower part of the face. In this work, a dual-input Neural Network architecture is proposed. The structure is a Siamese-like model, with two identical parallel streams (called base models) that process the two inputs separately. The input is represented by RGB images of the right eye and the left eye belonging to the same subject. The outputs of the two base models are merged through a fusion layer. The aim is to investigate how deep feature aggregation affects periocular recognition. The experimentation is performed on the Masked Face Recognition Database (M 2 FRED) which includes videos of 46 participants with and without masks. Three different fusion layers are applied to understand which type of merging technique is most suitable for data aggregation. Experimental results show promising performance for almost all experimental configurations with a worst-case accuracy of 90% and a best-case accuracy of 97%. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Relation Classification Based on Vietnamese Covid-19 Information Using BERT Model with Typed Entity Markers

This paper presents a study using the Bidirectional Encoder Representations from Transformers (BERT) base model to classifying relations based on Vietnamese Covid-19 information. The study applies two BERT-base models: R-BERT and BERT with entity start. In this work, instead of using entity markers for input, typed entity markers are used. The typed entities include the patient with name, the patient with age, the patient with the job, patient with gender, patient with symptom and disease, patient with transportation. A Vietnamese dataset is labeled manually and the final Bert base model to classify Covid-19 relation is slightly better than the model applied entity marked. © 2021, Springer Nature Singapore Pte Ltd.