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.
An electrical resistance sensor-based atmospheric corrosion monitor was employed to study the carbon steel corrosion in outdoor atmospheric environments by recording dynamic corrosion data in real-time. Data mining of collected data contributes to uncovering the underlying mechanism of atmospheric corrosion. In this study, it was found that most statistical correlation coefficients do not adapt to outdoor coupled corrosion data. In order to deal with online coupled data, a new machine learning model is proposed from the viewpoint of information fusion. It aims to quantify the contribution of different environmental factors to atmospheric corrosion in different exposure periods. Compared to the commonly used machine learning models of artificial neural networks and support vector machines in the corrosion research field, the experimental results demonstrated the efficiency and superiority of the proposed model on online corrosion data in terms of measuring the importance of atmospheric factors and corrosion prediction accuracy.
For semi-supervised learning, we propose the label propagation method in the framework of evidence theory. The proposed method can deal with two problems which are the classification tasks with extremely few labeled data and identification of outlier samples. By incorporating the basic belief assignment functions as the soft evidential labels, both the uncertainty of samples’ class assignments and the degree of samples’ abnormality can be measured. On the basis of the designed label update mechanism, the soft evidential label of each initially unlabeled sample will be iteratively updated by absorbing the label information from its neighbors. Once the partition of the dataset is stable, the samples with high confidences of abnormality will be identified as outliers and others will be assigned to their most supported classes. Experiment results on several UCI and benchmark datasets validate the effectiveness of the proposed method in comparison with several state-of-the-art methods. And tests on noisy spiral datasets show the robustness and feasibility of detecting outliers.
Uncertainty estimation is vital for machine learning with scientific data. Although existing methods effectively address inherent model uncertainty, they frequently overlook explicit modeling of complex noise in data devoid of temporal or spatial dependencies. This gap is especially challenging in structured scientific data, where such dependencies are commonly lacking. To address these challenges in scientific research, we propose the Taylor-Sensus Network (TSNet). TSNet innovatively uses a Taylor series expansion to model complex, heteroscedastic noise and proposes a deep Taylor block for aware noise distribution. TSNet includes a noise-aware contrastive learning module and a data density perception module for aleatoric and epistemic uncertainty. Additionally, an uncertainty combination operator is used to integrate these uncertainties, and the network is trained using a novel heteroscedastic mean square error loss. TSNet demonstrates superior performance over mainstream and state-of-the-art methods in experiments, highlighting its potential in scientific research and noise resistance.