A Federated Attention-Based Multimodal Biometric Recognition Approach in IoT.

The rise of artificial intelligence applications has led to a surge in Internet of Things (IoT) research. Biometric recognition methods are extensively used in IoT access control due to their convenience. To address the limitations of unimodal biometric recognition systems, we propose an attention-based multimodal biometric recognition (AMBR) network that incorporates attention mechanisms to extract biometric features and fuse the modalities effectively. Additionally, to overcome issues of data privacy and regulation associated with collecting training data in IoT systems, we utilize Federated Learning (FL) to train our model This collaborative machine-learning approach enables data parties to train models while preserving data privacy. Our proposed approach achieves 0.68%, 0.47%, and 0.80% Equal Error Rate (EER) on the three VoxCeleb1 official trial lists, performs favorably against the current methods, and the experimental results in FL settings illustrate the potential of AMBR with an FL approach in the multimodal biometric recognition scenario.

Knowledge Transfer via Pre-training for Recommendation: A Review and Prospect.

Recommender systems aim to provide item recommendations for users and are usually faced with data sparsity problems (e.g., cold start) in real-world scenarios. Recently pre-trained models have shown their effectiveness in knowledge transfer between domains and tasks, which can potentially alleviate the data sparsity problem in recommender systems. In this survey, we first provide a review of recommender systems with pre-training. In addition, we show the benefits of pre-training to recommender systems through experiments. Finally, we discuss several promising directions for future research of recommender systems with pre-training. The source code of our experiments will be available to facilitate future research.

Role of the CHADS2 Score in the Evaluation of Carotid Atherosclerosis in Patients with Atrial Fibrillation Undergoing Carotid Artery Ultrasonography.

OBJECTIVE: This study investigated the characteristics of carotid atherosclerosis in patients with atrial fibrillation (AF) and determined the feasibility and significance of the CHADS2 score in predicting the degree of carotid atherosclerosis. METHODS: Consecutive patients (n = 109) with nonvalvular AF were registered and classified into two groups, the paroxysmal AF group (n = 59) and persistent AF group (n = 50). Fifty healthy patients, matched by sex and age, were considered the control group. All patients were examined using carotid ultrasound and velocity vector imaging (VVI). RESULTS: Compared with the control group, the mean intimal-medial thickness in the paroxysmal AF group (0.56 ± 0.11 versus 0.61 ± 0.10, respectively, P < 0.05) and the persistent AF group (0.56 ± 0.11 versus 0.64 ± 0.13, respectively, P < 0.001) was significantly increased. The plaque index (PI) in the persistent AF group was significantly higher than that observed in the paroxysmal AF group (1.05 ± 1.33 versus 1.42 ± 1.47, respectively, P < 0.001). Regarding the VVI indices, those reflecting the long-axis longitudinal motion function of carotid arteries were significantly decreased in both AF groups. Compared with the control group, a significantly lower total longitudinal displacement (tLoD) index was observed in the persistent AF group (0.73 ± 0.66 versus 0.31 ± 0.23, respectively, P < 0·0001) and the paroxysmal AF group (0.73 ± 0.66 versus 0.34 ± 0.17, P < 0·0001). The CHADS2 score was related to indicators reflecting the structure and function of the carotid artery. CONCLUSIONS: Carotid arterial structure and function were significantly altered in patients with AF. The degree of carotid atherosclerosis depended on the duration of AF. The CHADS2 score may be useful as a predictor of the extent of carotid atherosclerosis in patients with AF.

Relationship of CHA2DS2-VASc and CHADS2 score to left atrial remodeling detected by velocity vector imaging in patients with atrial fibrillation.

BACKGROUND: The CHADS2/CHA2DS2-VASc scores are used to predict thrombo-embolic/stroke in patients with nonvalvular atrial fibrillation (AF). Nevertheless, limited data are available regarding the association between these risk stratification for stroke and left atrial (LA) remodeling status of AF patients. The purpose of this study was to explore the association between these scores and LA remodeling status assessed quantificationally by echocardiography in AF patients. METHODS: One hundred AF patients were divided into 3 groups based on the CHA2DS2-VASc/CHADS2 score: the score of 0 (low stroke risk), the score of 1 (moderate stroke risk) and the score of ≥2 (high stroke risk). All patients were performed through conventional and velocity vector imaging echocardiography. Echocardiographic parameters: maximum LA volume index (LAVImax), LA total emptying fraction (LAEFt) and LA mean strain were obtained to assess quantificationally LA remodeling status. RESULTS: On categorizing with CHA2DS2-VASc, the score of 1 group showed augment in LAVImax and attenuation in LA mean strain derived from VVI, compared with the score of 0 group (LAVImax: 40.27±21.91 vs. 26.79±7.87, p=0.002; LA mean strain: 15.18±6.36 vs. 22±8.54, p=0.001). On categorizing with the CHADS2 score, similar trends were seen between the score of ≥2 and 1 groups (LAVImax: 43.72±13.77 vs. 31.41±9.50, p<0.001; LA mean strain: 11.01±5.31 vs. 18.63±7.00, p<0.001). With multivariate logistic regression, LAVImax (odds ratio: 0.92 , 95% C=I: 0.85 to 0.98, p= 0.01) and LA mean strain reflecting LA remodeling (odds ratio: 1.10, 95% CI: 1.02 to 1.19, p=0.01) were strongly predictive of the CHA2DS2-VASc score of 0. CONCLUSIONS: The superiority of the CHADS2 score may lay in identifying LA remodeling of AF patients with high stroke risk. Whereas, the CHA2DS2-VASc score was better than the CHADS2 score at identifying LA remodeling of AF patients presenting low stroke risk.