Automating dysarthria assessments offers the opportunity to develop practical, low-cost tools that address the current limitations of manual and subjective assessments. Nonetheless, the small size of ...most dysarthria datasets makes it challenging to develop automated assessment. Recent research showed that speech representations from models pre-trained on large unlabelled data can enhance Automatic Speech Recognition (ASR) performance for dysarthric speech. We are the first to evaluate the representations from pre-trained state-of-the-art Self-Supervised models across three downstream tasks on dysarthric speech: disease classification, word recognition and intelligibility classification, and under three noise scenarios on the UA-Speech dataset. We show that HuBERT is the most versatile feature extractor across dysarthria classification, word recognition, and intelligibility classification, achieving respectively \(+24.7\%, +61\%, \text{and} +7.2\%\) accuracy compared to classical acoustic features.
ObjectivesTo quantify the associations between shielding status and loneliness at the start of the COVID-19 pandemic, and physical activity (PA) levels throughout the pandemic.MethodsDemographic, ...health and lifestyle characteristics of 7748 cognitively healthy adults aged >50, and living in London, were surveyed from April 2020 to March 2021. The International Physical Activity Questionnaire (IPAQ) short-form assessed PA before COVID-19 restrictions, and up to 6 times over 11 months. Linear mixed models investigated associations between shielding status and loneliness at the onset of the pandemic, with PA over time.ResultsParticipants who felt 'often lonely' at the outset of the pandemic completed an average of 522 and 547 fewer Metabolic Equivalent of Task (MET) minutes/week during the pandemic (95% CI: -809, -236, p<0.001) (95% CI: -818, -275, p<0.001) than those who felt 'never lonely' in univariable and multivariable models adjusted for demographic factors respectively. Those who felt 'sometimes lonely' completed 112 fewer MET minutes/week (95% CI: -219, -5, p = 0.041) than those who felt 'never lonely' following adjustment for demographic factors. Participants who were shielding at the outset of the pandemic completed an average of 352 fewer MET minutes/week during the pandemic than those who were not (95% CI: -432, -273; p<0.001) in univariable models and 228 fewer MET minutes/week (95% CI: -307, -150, p<0.001) following adjustment for demographic factors. No significant associations were found after further adjustment for health and lifestyle factors.ConclusionsThose shielding or lonely at pandemic onset were likely to have completed low levels of PA during the pandemic. These associations are influenced by co-morbidities and health status.
Errors might not have the same consequences depending on the task at hand. Nevertheless, there is limited research investigating the impact of imbalance in the contribution of different features in ...an error vector. Therefore, we propose the Feature Impact Balance (FIB) score. It measures whether there is a balanced impact of features in the discrepancies between two vectors. We designed the FIB score to lie in 0, 1. Scores close to 0 indicate that a small number of features contribute to most of the error, and scores close to 1 indicate that most features contribute to the error equally. We experimentally study the FIB on different datasets, using AutoEncoders and Variational AutoEncoders. We show how the feature impact balance varies during training and showcase its usability to support model selection for single output and multi-output tasks.
Despite impressive empirical advances of SSL in solving various tasks, the problem of understanding and characterizing SSL representations learned from input data remains relatively under-explored. ...We provide a comparative analysis of how the representations produced by SSL models differ when masking parts of the input. Specifically, we considered state-of-the-art SSL pretrained models, such as DINOv2, MAE, and SwaV, and analyzed changes at the representation levels across 4 Image Classification datasets. First, we generate variations of the datasets by applying foreground and background segmentation. Then, we conduct statistical analysis using Canonical Correlation Analysis (CCA) and Centered Kernel Alignment (CKA) to evaluate the robustness of the representations learned in SSL models. Empirically, we show that not all models lead to representations that separate foreground, background, and complete images. Furthermore, we test different masking strategies by occluding the center regions of the images to address cases where foreground and background are difficult. For example, the DTD dataset that focuses on texture rather specific objects.
This paper proposes the position control of DC motor. Two methods are used for position control, LQR method and feedback linearization. We show that these methods without load torque are stable, but, ...when load is added to the motor's shaft, LQR and feedback linearization can not make efficient input signal for reference tracking in output. To solve this problem, we combined these methods and will show by using combined method, the position of shaft tracks reference in presence of large torque. For validation of new controller, we compared response with LQR and feedback linearization. Simulation results show stable response of new method.