A Summary of the ComParE COVID-19 Challenges (preprint)

The COVID-19 pandemic has caused massive humanitarian and economic damage. Teams of scientists from a broad range of disciplines have searched for methods to help governments and communities combat the disease. One avenue from the machine learning field which has been explored is the prospect of a digital mass test which can detect COVID-19 from infected individuals' respiratory sounds. We present a summary of the results from the INTERSPEECH 2021 Computational Paralinguistics Challenges: COVID-19 Cough, (CCS) and COVID-19 Speech, (CSS).

Capturing the experience and lessons from junior doctors working at the Nightingale North West: a qualitative study

NHS Nightingale Hospital North West (NNW) was a new temporary hospital within the NHS designed to rapidly expand capacity to care for patients during the COVID-19 pandemic. Within 2 weeks, Manchester Central Convention Centre was converted into a potential 648 bed facility, capable of providing step-down care to patients from the north west.Junior doctors had the opportunity to witness the creation of a field hospital, shape systems and processes, and work with a diverse team coming together for a common cause. To capture their experiences, interviews were conducted using a semi-structured format and the responses summarised into transcripts. Consensus coding was performed using domains/themes.When exploring successes, there was consistent mention of a strong team;in particular the feeling of being individually valued within a flattened hierarchy. Staff wellbeing and education were also regularly mentioned and helped contribute to this overall feeling. When asked what they would take forward, doctors focussed on the importance of a strong team that values multi-disciplinary working.But the hospital was not without challenges, with processes changing from one shift to the next and leading to potential errors. In addition, system issues (such as with medication and documentation) lead to a sometimes-chaotic work environment. Staff identification was a significant challenge, and potentially contributed to communication breakdowns.To rectify this, doctors undertook QI projects which formed the basis for re-activation plans. Perhaps more important than material improvements were feelings of empowerment they identified to achieve actionable change within the hospital.Junior doctors were overwhelmingly positive about their NNW experience. Their power to act as agents of change was showcased at NNW, where senior management encouraged them to take ownership of challenges identified and seek ways to improve the system in which they worked.

The INTERSPEECH 2021 Computational Paralinguistics Challenge: COVID-19 Cough, COVID-19 Speech, Escalation & Primates (preprint)

The INTERSPEECH 2021 Computational Paralinguistics Challenge addresses four different problems for the first time in a research competition under well-defined conditions: In the COVID-19 Cough and COVID-19 Speech Sub-Challenges, a binary classification on COVID-19 infection has to be made based on coughing sounds and speech; in the Escalation SubChallenge, a three-way assessment of the level of escalation in a dialogue is featured; and in the Primates Sub-Challenge, four species vs background need to be classified. We describe the Sub-Challenges, baseline feature extraction, and classifiers based on the 'usual' COMPARE and BoAW features as well as deep unsupervised representation learning using the AuDeep toolkit, and deep feature extraction from pre-trained CNNs using the Deep Spectrum toolkit; in addition, we add deep end-to-end sequential modelling, and partially linguistic analysis.

End-2-End COVID-19 Detection from Breath & Cough Audio (preprint)

Our main contributions are as follows: (I) We demonstrate the first attempt to diagnose COVID-19 using end-to-end deep learning from a crowd-sourced dataset of audio samples, achieving ROC-AUC of 0.846; (II) Our model, the COVID-19 Identification ResNet, (CIdeR), has potential for rapid scalability, minimal cost and improving performance as more data becomes available. This could enable regular COVID-19 testing at apopulation scale; (III) We introduce a novel modelling strategy using a custom deep neural network to diagnose COVID-19 from a joint breath and cough representation; (IV) We release our four stratified folds for cross parameter optimisation and validation on a standard public corpus and details on the models for reproducibility and future reference.