Search

Mastication plays an important role in effective food digestion and nutrient absorption. Therefore, regulating masticatory force in people with declining mastication function is significant for maintaining health and quality of life. In this study, we tested the effect of tactile augmentation on mastication force. To augment tactile feedback during mastication, we applied closed-loop electrical stimulation onto the mandibular vestibule using an intraoral tooth-borne electronic system. We hypothesized that closed-loop electrical stimulation, timed with mastication and applied to the nerves delivering tactile feedback to the brain, would evoke an increase in masticatory force. Experiments were completed using the intraoral system with six healthy human subjects who masticated soft and hard foods with and without stimulation during the experiment. Their mastication forces were recorded ten times per condition. The recorded mastication force profile showed that mastication force was higher with the harder food. Also, mastication force increased when electrical stimulation was applied, compared to the non-stimulated condition. These results support the hypothesis that tactile augmentation by intraoral closed-loop electrical stimulation will increase masticatory force. Other mastication parameters including period, spike width, and duty cycle are also changed by electrical stimulation. Further, stimulation left a strong aftereffect on these mastication parameters.

Privacy protection is considered as an important problem in learning-based systems. Recently, various works based on differential privacy have been proposed to protect an individual's privacy in the machine learning and deep learning contexts. One of the state-of-the-art approaches is Private Aggregation of Teacher Ensembles (PATE), a generic framework which can be successfully applied to many different learning algorithms. In PATE, we need to split the private dataset into many disjoint subsets and train an ensemble of teachers on these subsets. Then, we transfer noisy predictions from the ensemble of teachers to a student model. In this paper, we show that for complex datasets and tasks, such as nature image classification, the training set allocated for one teacher may be too small with respect to the corresponding task to achieve an ideal performance. To alleviate this problem, we propose the TrPATE framework which extends PATE with transfer learning. Based on PATE, we transfer and share the knowledge extracted from a publicly available non-private dataset to the teachers. The extensive experiments are conducted on various datasets, and the empirical results demonstrate the effectiveness of our method.