Deep learning framework for material design space exploration using active transfer learning and data augmentation

Neural network-based generative models have been actively investigated as an inverse design method for finding novel materials in a vast design space. However, the applicability of conventional generative models is limited because they cannot access data outside the range of training sets. Advanced generative models that were devised to overcome the limitation also suffer from the weak predictive power on the unseen domain. In this study, we propose a deep neural network-based forward design approach that enables an efficient search for superior materials far beyond the domain of the initial training set. This approach compensates for the weak predictive power of neural networks on an unseen domain through gradual updates of the neural network with active transfer learning and data augmentation methods. We demonstrate the potential of our framework with a grid composite optimization problem that has an astronomical number of possible design configurations. Results show that our proposed framework can provide excellent designs close to the global optima, even with the addition of a very small dataset corresponding to less than 0.5% of the initial training dataset size.

Deep Learning Framework for Material Design Space Exploration using Active Transfer Learning and Data Augmentation

Neural network-based generative models have been actively investigated as an inverse design method for finding novel materials in a vast design space. However, the applicability of conventional generative models is limited because they cannot access data outside the range of training sets. Advanced generative models that were devised to overcome the limitation also suffer from the weak predictive power on the unseen domain. In this study, we propose a deep neural network-based forward design approach that enables an efficient search for the superior materials far beyond the domain of the initial training set. This approach compensates for the weak predictive power of neural networks on an unseen domain through gradual updates of the neural network with active-transfer learning and data augmentation methods. We demonstrate the potential of our framework with a grid composite optimization problem that has an astronomical number of possible design configurations. Results show that our proposed framework can provide excellent designs close to the global optima, even with the addition of very small dataset corresponding to less than 0.5% of the initial training dataset size.

Study on Human Activity Recognition Using Semi-Supervised Active Transfer Learning

In recent years, various studies have begun to use deep learning models to conduct research in the field of human activity recognition (HAR). However, there has been a severe lag in the absolute development of such models since training deep learning models require a lot of labeled data. In fields such as HAR, it is difficult to collect data and there are high costs and efforts involved in manual labeling. The existing methods rely heavily on manual data collection and proper labeling of the data, which is done by human administrators. This often results in the data gathering process often being slow and prone to human-biased labeling. To address these problems, we proposed a new solution for the existing data gathering methods by reducing the labeling tasks conducted on new data based by using the data learned through the semi-supervised active transfer learning method. This method achieved 95.9% performance while also reducing labeling compared to the random sampling or active transfer learning methods.

Fetal maternal immune cross talking and vaccination during pregnancy

Maternal adaptations in pregnancy induce complex physiological changes of the immune system which protect mother’s health and ensure the accommodation of the growing embryo. The innate immunity is amplified and the adaptive immunity is partially suppressed, preserving the ability to produce antibodies. Vaccination during pregnancy constitutes a fundamental preventive measure in the antenatal care, as it achieves both the immunization of the mother and the baby. Besides passive immunization through the placenta, also active transfer of IgG via placenta occurs throughout pregnancy. Live attenuated vaccines (LAV) are contraindicated in pregnant women, while the others should be recommended if benefits overshadow risks.

Trogocytosis between Non-Immune Cells for Cell Clearance, and among Immune-Related Cells for Modulating Immune Responses and Autoimmunity

The term trogocytosis refers to a rapid bidirectional and active transfer of surface membrane fragment and associated proteins between cells. The trogocytosis requires cell-cell contact, and exhibits fast kinetics and the limited lifetime of the transferred molecules on the surface of the acceptor cells. The biological actions of trogocytosis include information exchange, cell clearance of unwanted tissues in embryonic development, immunoregulation, cancer surveillance/evasion, allogeneic cell survival and infectious pathogen killing or intercellular transmission. In the present review, we will extensively review all these aspects. In addition to its biological significance, aberrant trogocytosis in the immune system leading to autoimmunity and immune-mediated inflammatory diseases will also be discussed. Finally, the prospective investigations for further understanding the molecular basis of trogocytosis and its clinical applications will also be proposed.

Tyrosol-Enriched Tomatoes by Diffusion across the Fruit Peel from a Chitosan Coating: A Proposal of Functional Food

Chitosan is receiving increasing attention from the food industry for being a biodegradable, non-toxic, antimicrobial biopolymer able to extend the shelf life of, and preserve the quality of, fresh food. However, few studies have investigated the ability of chitosan-based coatings to allow the diffusion of bioactive compounds into the food matrix to improve its nutraceutical quality. This research is aimed at testing whether a hydrophilic molecule (tyrosol) could diffuse from the chitosan-tyrosol coating and cross the tomato peel. To this end, in vitro permeation tests using excised tomato peel and an in vivo application of chitosan-tyrosol coating on tomato fruit, followed by tyrosol quantification in intact fruit, peel and flesh during a seven-day storage at room temperature, were performed. Both approaches demonstrated the ability of tyrosol to permeate across the fruit peel. Along with a decreased tyrosol content in the peel, its concentration within the flesh was increased, indicating an active transfer of tyrosol into this tissue. This finding, together with the maintenance of constant tyrosol levels during the seven-day storage period, is very promising for the use of chitosan formulations to produce functional tomato fruit.