Transfer learning for the efficient detection of COVID-19 from smartphone audio data.

Disease detection from smartphone data represents an open research challenge in mobile health (m-health) systems. COVID-19 and its respiratory symptoms are an important case study in this area and their early detection is a potential real instrument to counteract the pandemic situation. The efficacy of this solution mainly depends on the performances of AI algorithms applied to the collected data and their possible implementation directly on the users' mobile devices. Considering these issues, and the limited amount of available data, in this paper we present the experimental evaluation of 3 different deep learning models, compared also with hand-crafted features, and of two main approaches of transfer learning in the considered scenario: both feature extraction and fine-tuning. Specifically, we considered VGGish, YAMNET, and L3-Net (including 12 different configurations) evaluated through user-independent experiments on 4 different datasets (13,447 samples in total). Results clearly show the advantages of L3-Net in all the experimental settings as it overcomes the other solutions by 12.3% in terms of Precision-Recall AUC as features extractor, and by 10% when the model is fine-tuned. Moreover, we note that to fine-tune only the fully-connected layers of the pre-trained models generally leads to worse performances, with an average drop of 6.6% with respect to feature extraction. Finally, we evaluate the memory footprints of the different models for their possible applications on commercial mobile devices.

Caring more about food: The unexpected positive effect of the Covid-19 lockdown on household food management and waste.

Over half of the total amount of food wasted in Europe concerns household food waste which is mainly due to incorrect food management habits and behaviour. During the Covid-19 outbreak, food management and consumption habits changed dramatically due to the tough lockdown restrictions imposed by governments to reduce infection. This study investigated how these dramatic changes in the daily lives of consumers influenced the generation of food waste at household level. A CAWI questionnaire was administered to a sample of 1078 Italian consumers during the lockdown (March-April 2020). The respondents were asked to self-estimate the percentage of food their households wasted before and during the lockdown and to explain their food management habits. We focused the analysis on the differences between the food the respondents declared to have wasted before and during lockdown, which revealed that most households threw away less food during the Covid-19 lockdown compared to the pre-Covid situation. We referred to Seemingly Unrelated Regression models to evaluate the association between the food waste behaviour in the two periods considered in the study and the other factors observed. The results disclosed that young consumers and people who started implementing good food management practices (shopping list, meal planning etc.) more frequently considerably reduced the food they wasted during lockdown. Also, the logistical difficulties of grocery shopping experienced by consumers during lockdown made them manage their household food consumption more carefully, which led to a reduction in the amount of food wasted.

Speech and Swallowing Outcomes Following Surgical Resection with Immediate Free Tissue Transfer Reconstruction for Advanced Osteoradionecrosis of the Mandible Following Radiation Treatment for Head and Neck Cancer.

Despite recent advances in the radiation techniques used for the treatment of head and neck cancer (HNC) including intensity-modulated radiotherapy (IMRT), mandibular osteoradionecrosis (ORN) remains a significant complication. Advanced stage ORN is managed surgically with resection and immediate free tissue transfer reconstruction. An evaluation of the functional speech and swallowing outcomes was undertaken for patients undergoing surgical management of advanced ORN. We retrospectively reviewed consecutive patients, at a single, tertiary cancer centre, who underwent surgical resection for advanced Notani grade III ORN. Outcomes investigated included use and duration of tracheostomy and swallowing and speech status using Performance Status Scale for Head and Neck Cancer Normalcy of Diet (PSS-NOD) and Understandability of Speech (PSS-Speech) at baseline and 3 months following surgery. Ten patients underwent surgical resection with free tissue transfer reconstruction between January 2014 and December 2019. Two patients required supplemental nutrition via a gastrostomy at three months post surgery. As per the PSS-NOD data half of the patients' (n = 5) diet remained stable (n = 2) or improved (n = 3) and half of the participants experienced a decline in diet (n = 5). The majority of patients had no speech difficulties at baseline (n = 8). The majority of patients' speech remained stable (n = 8) with two patients experiencing a deterioration in speech clarity following surgery. Well-designed studies with robust, sensitive multidimensional dysphagia and communication assessments are required to fully understand the impact of surgical management of advanced ORN using resection with free tissue transfer reconstruction.

ContextLabeler dataset: Physical and virtual sensors data collected from smartphone usage in-the-wild.

This paper describes a data collection campaign and the resulting dataset derived from smartphone sensors characterizing the daily life activities of 3 volunteers in a period of two weeks. The dataset is released as a collection of CSV files containing more than 45K data samples, where each sample is composed by 1332 features related to a heterogeneous set of physical and virtual sensors, including motion sensors, running applications, devices in proximity, and weather conditions. Moreover, each data sample is associated with a ground truth label that describes the user activity and the situation in which she was involved during the sensing experiment (e.g., working, at restaurant, and doing sport activity). To avoid introducing any bias during the data collection, we performed the sensing experiment in-the-wild, that is, by using the volunteers' devices, and without defining any constraint related to the user's behavior. For this reason, the collected dataset represents a useful source of real data to both define and evaluate a broad set of novel context-aware solutions (both algorithms and protocols) that aim to adapt their behavior according to the changes in the user's situation in a mobile environment.