Power Efficient Machine Learning Models Deployment on Edge IoT Devices.

Computing has undergone a significant transformation over the past two decades, shifting from a machine-based approach to a human-centric, virtually invisible service known as ubiquitous or pervasive computing. This change has been achieved by incorporating small embedded devices into a larger computational system, connected through networking and referred to as edge devices. When these devices are also connected to the Internet, they are generally named Internet-of-Thing (IoT) devices. Developing Machine Learning (ML) algorithms on these types of devices allows them to provide Artificial Intelligence (AI) inference functions such as computer vision, pattern recognition, etc. However, this capability is severely limited by the device's resource scarcity. Embedded devices have limited computational and power resources available while they must maintain a high degree of autonomy. While there are several published studies that address the computational weakness of these small systems-mostly through optimization and compression of neural networks- they often neglect the power consumption and efficiency implications of these techniques. This study presents power efficiency experimental results from the application of well-known and proven optimization methods using a set of well-known ML models. The results are presented in a meaningful manner considering the "real world" functionality of devices and the provided results are compared with the basic "idle" power consumption of each of the selected systems. Two different systems with completely different architectures and capabilities were used providing us with results that led to interesting conclusions related to the power efficiency of each architecture.

Epigenetic/genetic variations in CG-rich elements of immune-related genes contribute to food allergy development during childhood.

BACKGROUND: Genetic areas of FOXP3 TSDR, human leukocyte antigen-G (HLA-G) upstream of CpG island 96, CpG41 and CpG73 islands of the HLA-DRB1 and HLA-DQB1 genes respectively, previously documented to display immune-modulatory properties, were subjected to epigenetic/genetic analysis to assess their influence in IgE-mediated food allergy (FA) development in children. METHODS: Sixty-four orally challenged and IgE-tested food allergic subjects together with 44 controls were recruited. Targeted pyrosequencing analysis to detect DNA methylation status and genetic variations was utilized and experimental results obtained were analyzed by a statistical software platform and correlated to clinical data. Also, transcription factor (TF) binding sites in study areas were unmasked by the JASPAR prediction database. RESULTS: Parents' smoking was significantly correlated with aberrant methylation patterns, regardless of food allergic or control status. HLA-G promoter region showed a trend for hypomethylation in food allergic subjects, with one of the CG sites displaying significantly decreased methylation values. Rs1233333, residing within the HLA-G promoter region preserved a protective role toward DNA methylation. Variable methylation patterns were recorded for CpG41 of the HLA-DRB1 gene and hypermethylation of the region was significantly correlated with the presence of single nucleotide polymorphisms (SNPs). TFs' recognition sites, located in studied genetic areas and exerting pivotal regulatory biological roles, are potentially affected by divergent DNA methylation status. CONCLUSIONS: We propose that HLA-G expression is triggered by food-derived allergens, providing a TregFoxP3-/HLA-G+ subpopulation generation to promote direct immune tolerance. Furthermore, clear evidence is provided for the underlying co-operation of genetic polymorphisms with epigenetic events, mainly at the CpG41 island of the HLA-DRB1 gene, which needs an extended investigation and elucidation.

Association of genome variations in the renin-angiotensin system with physical performance.

BACKGROUND: The aim of this study was to determine the genotype distribution and allelic frequencies of ACE (I/D), AGTR1 (A +1166 C), BDKRB2 (+9/-9) and LEP (G-2548A) genomic variations in 175 Greek athletes who excelled at a national and/or international level and 169 healthy Greek adults to identify whether some particular combinations of these loci might serve as predictive markers for superior physical condition. RESULTS: The D/D genotype of the ACE gene (p = 0.034) combined with the simultaneous existence of BDKRB2 (+9/-9) (p = 0.001) or LEP (G/A) (p = 0.021) genotypes was the most prevalent among female athletes compared to female controls. A statistical trend was also observed in BDKRB2 (+9/-9) and LEP (G-2548A) heterozygous genotypes among male and female Greek athletes, and in ACE (I/D) only in male athletes. Finally, both male and female athletes showed the highest rates in the AGTR1 (A/A) genotype. CONCLUSIONS: Our results suggest that the co-existence of ACE (D/D), BDKRB2 (+9/-9) or LEP (G/A) genotypes in female athletes might be correlated with a superior level of physical performance.

Authentication and self-correction in sequential MRI slices.

One of the new challenges of Information Technology in the medical world is the protection and authentication of a variety of digital medical files, datasets, and images. In this work, the ability of magnetic resonance imaging (MRI) slice sequences to hide digital data is investigated and more specifically the case that the hidden data are the regions of interest (ROI) of the MRI slices. The regions of non-interest (RONI) are used as cover. The hiding capacity of the whole sequence is taken into account. Any ROI-targeted tampering attempt can be detected, and the original image can be self-restored (under certain conditions) by extracting the ROI from the RONI.