A multimodal deep learning architecture for smoking detection with a small data approach.

Covert tobacco advertisements often raise regulatory measures. This paper presents that artificial intelligence, particularly deep learning, has great potential for detecting hidden advertising and allows unbiased, reproducible, and fair quantification of tobacco-related media content. We propose an integrated text and image processing model based on deep learning, generative methods, and human reinforcement, which can detect smoking cases in both textual and visual formats, even with little available training data. Our model can achieve 74% accuracy for images and 98% for text. Furthermore, our system integrates the possibility of expert intervention in the form of human reinforcement. Using the pre-trained multimodal, image, and text processing models available through deep learning makes it possible to detect smoking in different media even with few training data.

Identification of high radon dwellings, risk of exposure, and geogenic potential in the mining area of the "TREPCA" complex.

This study was performed to investigate radon levels in dwellings in the mining area near the town of Kosovska Mitrovica. The Passive radon technique based on the charcoal canister test kit conducted in summer and autumn 2019 showed unexpected results. The reference level of 300 Bq/m3 for indoor radon concentration was exceeded in 15 of 26 dwellings. Preliminary measurements of gamma dose rate in some dwellings built from local stone showed values from 0.30 to 0.45 µSv/h, while 75% of measurements in dwellings with stone foundations had radon above the reference level. The highest radon concentration (22 500 ± 220 Bq/m3 ) was measured in the cellar of one family house. The RAD7 device (Durridge Company, Inc.) was used to measure radon concentrations in water and nearby soil of some dwellings. Indoor radon concentration fluctuated significantly over two days; over a one-day time scale, radon varied from 2843 ± 217 Bq/m3 at midnight to 1449 ± 104 Bq/m3 in the morning at one site, and abruptly decreased from a maximum of 2146 ± 262 Bq/m3 in one day to a minimum of 21 Bq/m3 the next day at another site. The influence of geological substrate on radon exposure was discussed through the estimation of geogenic radon potential, which varies from low the high radon index despite to high permeability of soil.

Improved non-destructive method for 90Sr activity determination in aqueous solutions using Monte Carlo simulation.

The development of new methods and improvements of existing methods for the specific activity determination of 90Sr and other distinct beta emitters has been of considerable interest. The reason for this interest is that the notably small number of methods that are able to meet all the set criteria, such as reliability of the results, measurement uncertainty and time, and minimum production of radioactive waste, as well as applicability to various samples with reference to their nature, geometry and composition. In this paper, two methods for rapid 90Sr activity determination based on Monte Carlo simulations are used, one for a Si semiconductor detector for beta spectrometric measurements and the other for the Geiger-Muller (GM) ionization probe. To improve the reliability of the measurement results, samples with high and low strontium activity solutions were prepared in the form of dry residues. The results of the proposed methodology were verified with a standard method using a liquid scintillation counter, and notably good agreements are achieved.