Case-base maintenance with multi-objective evolutionary algorithms

Case-Base Reasoning is a problem-solving methodology that uses old solved problems, called cases, to solve new problems. The case-base is the knowledge source where the cases are stored, and the amount of stored cases is critical to the problem-solving ability of the Case-Base Reasoning system. However, when the case-base has many cases, then performance problems arise due to the time needed to find those similar cases to the input problem. At this point, Case-Base Maintenance algorithms can be used to reduce the number of cases and maintain the accuracy of the Case-Base Reasoning system at the same time. Whereas Case-Base Maintenance algorithms typically use a particular heuristic to remove (or select) cases from the case-base, the resulting maintained case-base relies on the proportion of redundant and noisy cases that are present in the case-base, among other factors. That is, a particular Case-Base Maintenance algorithm is suitable for certain types of case-bases that share some indicators, such as redundancy and noise levels. In the present work, we consider Case-Base Maintenance as a multi-objective optimization problem, which is solved with a Multi-Objective Evolutionary Algorithm. To this end, a fitness function is introduced to measure three different objectives based on the Complexity Profile model. Our hypothesis is that the Multi-Objective Evolutionary Algorithm performing Case-Base Maintenance may be used in a wider set of case-bases, achieving a good balance between the reduction of cases and the problem-solving ability of the Case-Based Reasoning system. Finally, from a set of the experiments, our proposed Multi-Objective Evolutionary Algorithm performing Case-Base Maintenance shows regularly good results with different sets of case-bases with different proportion of redundant and noisy cases.     

R2D2: A scalable deep learning toolkit for medical imaging segmentation

Deep learning has gained a significant popularity in recent years thanks to its tremendous success across a wide range of relevant fields of applications, including medical image analysis domain in particular. Although convolutional neural networks (CNNs) based medical applications have been providing powerful solutions and revolutionizing medicine, efficiently training of CNNs models is a tedious and challenging task. It is a computationally intensive process taking long time and rare system resources, which represents a significant hindrance to scientific research progress. In order to address this challenge, we propose in this article, R2D2, a scalable intuitive deep learning toolkit for medical imaging semantic segmentation. To the best of our knowledge, the present work is the first that aims to tackle this issue by offering a novel distributed versions of two well-known and widely used CNN segmentation architectures [ie, fully convolutional network (FCN) and U-Net]. We introduce the design and the core building blocks of R2D2. We further present and analyze its experimental evaluation results on two different concrete medical imaging segmentation use cases. R2D2 achieves up to 17.5× and 10.4× speedup than single-node based training of U-Net and FCN, respectively, with a negligible, though still unexpected segmentation accuracy loss. R2D2 offers not only an empirical evidence and investigates in-depth the latest published works but also it facilitates and significantly reduces the effort required by researchers to quickly prototype and easily discover cutting-edge CNN configurations and architectures.     

Computational color analysis of paintings for different artists of the XVI and XVII centuries

We have performed a computational color analysis of images of paintings for six master painters: Titian, Rubens, El Greco, Velázquez, Rembrandt, and Vermeer. These painters show the evolution from the renaissance to the baroque style. Different first and second‐order statistical parameters have been obtained and analyzed in order to fix which of them can be common for the different artists and which of them can be representative of a certain period of time or the evolution of the art. The firsts include the orientation and semi‐axes ratio of the ellipses that define the gamut in the chromaticity diagram and the dependencies with the frequency of the power of the Fourier transforms. Most differences among artists can be found in the volume and area of the gamut, the number of discernible colors which is greater for Titian, El Greco and Rubens, compared to Velázquez, Rembrandt and Vermeer, the average value of L* and the number of dark pixels.     

Effect of drying methods on the reactivity of chitosan towards Maillard reaction

The effect of drying methods on chitosan reactivity towards Maillard reaction during storage of dried chitosan–lactose systems was investigated. Two different structural forms of chitosan, scaffold and microspheres, were prepared. Then they were dried by lyophilization or using supercritical fluid technology (SF-CO₂) and stored with lactose under controlled temperature conditions (60 °C) and water activity (a_w = 0.65). The drying method produced slight modifications on the chitosan structure especially in samples submitted to SF-CO₂ treatment. Differences were more evident in the reactivity of chitosan–lactose systems during storage. Maillard reaction development was assessed by quantification of a new compound which originated from the acid hydrolysis of Amadori compounds (NFMD) resulting from chitosan–lactose interactions, and not detected in the acid hydrolysates of freshly prepared chitosan-lactose systems. Maillard reaction was influenced by different factors such as the manner of lactose addition, structure of chitosan (microspheres and scaffolds) and drying method. Lyophilized chitosan in the form of microspheres with lactose, presented the highest values of NFMD (428.45 mg/100 g of sample). The determination of this compound could be used as an index of the progress of Maillard reaction.     

Batch and Fed‐Batch Degradation of Enrofloxacin by the Fenton Process

Enrofloxacin is a synthetic second‐generation fluoroquinolone used as an antimicrobial agent exclusively in veterinary medicine. To simulate the treatment of wastewater contaminated by enrofloxacin, four‐day long fed‐batch runs were carried out according to the Fenton process with an enrofloxacin solution as model, to which hydrogen peroxide and ferrous ion were added twice a day. The residual enrofloxacin concentration was practically coincident to that detected at the end of the batch tests. Hydrogen peroxide was almost completely consumed after each feeding period, while the total organic carbon (TOC) concentration decreased gradually within three days, corresponding to a reduction > 58 %. From the third day on, the TOC falling rate was quite low. A yellow sludge settled due to the precipitation of both Fe(OH)₃ and a complex formed by ferric ion with adsorbed enrofloxacin and/or its oxidation products.     

Properties and Printability of Inkjet and Screen-Printed Silver Patterns for RFID Antennas

We report the modeling, and geometrical and electrical characterization, of inkjet and screen-printed patterns on different polymeric substrates for use as antennas in radio-frequency identification (RFID) applications. We compared the physical and electrical characteristics of two silver nanoparticle-based commercial inkjet-printable inks and one screen-printable silver paste, when deposited on polyimide (PI), polyethylene terephthalate (PET), and polyetherimide (PEI) substrates. First, the thickness of the inkjet-printed patterns was predicted by use of an analytical model based on printing conditions and ink composition. The predicted thickness was confirmed experimentally, and geometrical characterization of the lines was completed by measuring the root-mean-square roughness of the patterns. Second, direct-current electrical characterization was performed to identify the printing conditions yielding the lowest resistivity and sheet resistance. The minimum resistivity for the inkjet-printing method was 8.6 ± 0.8 μΩ cm, obtained by printing four stacked layers of one of the commercial inks on PEI, whereas minimum resistivity of 44 ± 7 μΩ cm and 39 ± 4 μΩ cm were obtained for a single layer of screen-printed ink on polyimide (PI) with 140 threads/cm mesh and 90 threads/cm mesh, respectively. In every case, these minimum values of resistivity were obtained for the largest tested thickness. Coplanar waveguide transmission lines were then designed and characterized to analyze the radio-frequency (RF) performance of the printed patterns; minimum transmission losses of 0.0022 ± 0.0012 dB/mm and 0.0016 ± 0.0012 dB/mm measured at 13.56 MHz, in the high-frequency (HF) band, were achieved by inkjet printing on PEI and screen printing on PI, respectively. At 868 MHz, in the ultra-high-frequency band, the minimum values of transmission loss were 0.0130 ± 0.0014 dB/mm for inkjet printing on PEI and 0.0100 ± 0.0014 dB/mm for screen printing on PI. Although the resistivity achieved is lower for inkjet printing than for screen printing, RF losses for inkjetted patterns were larger than for screen-printed patterns, because thicker layers were obtained by screen printing. Finally, several coil inductors for the HF band were also fabricated by use of both printing techniques, and were used as antennas for semi-passive smart RFID tags on plastic foil capable of measuring temperature and humidity.     

Medical knowledge management for specific hospital departments

Medical knowledge representation and management is concerned with how to organise the often vague clinical experience of medical staff required for computable models. However, few knowledge management and acquisition tools have entered routine use, since such tools are not perceived by physicians as part of the clinical information process. An attempt to partially solve this problem, we identify two key aspects of knowledge representation and management tasks. The first is to adopt a medical knowledge standardisation to provide a consistent terminology control and to simplify the integration between knowledge management tools and the health information system. The second is to establish an effective knowledge acquisition process in specific medical fields by adapting knowledge acquisition tools. Therefore, the main goal of this work is to define computational models and to design mechanisms for the effective acquisition and management of medical knowledge in real-life hospital departments. To this end, we analyse the representation of medical knowledge (based on deep-causal models) and the development of knowledge management tools (based on ontologies), integrated within the information processing activities of the clinical user. Finally, we illustrate its applicability in the Intensive Care Unit and Pediatry scenarios.     

Non-destructive Estimation of Mandarin Maturity Status Through Portable VIS-NIR Spectrophotometer

Sugar content is one of the most important quality attributes of citrus fruit, either for fresh or for processing market. Since sugars in citrus juice are highly correlated with total soluble solids (TSS) content, which can be determined easily even by the means of a hand refractometer, TSS is one of the most frequently used quality index. Since TSS can be measured only destructively, the results are representative only if carried out on large samples and do not allow classifying marketable fruit one by one according to their specific sugar content. Objective of this experiment was to assess possibility and limits of a non-destructive estimation of citrus fruits internal quality parameters (TSS and titratable acidity) presenting thick peel by the use of a spectrophotometric portable VIS-NIR system. Four hundred fruit of “Miho” satsuma and 150 fruit of “Page” tangelo were used. Each fruit was first subjected to spectrophotometric acquisition and soon after was juiced and TSS and titratable acidity (TA) determined. Partial least squares (PLS) regression analysis was applied for constructing a predictive model based on the spectral normalized response, constructing the model on a sub-sample and verifying the model (prediction test) on independent ones. The TA relative to Page mandarin was predicted in the test with an r = 0.88 and a standard error of prevision (SEP) coefficient of variability of 3.8% while the TSS scored an r = 0.85 and a SEP coefficient of variability equal to 4%. The TA of Miho mandarin was predicted in the test with an r = 0.81 and a SEP coefficient of Variability of 8.3% while the TSS scored an r = 0.84 and a SEP coefficient of variability equal to 5.6%.