Techniques for Selecting the Optimal Parameters of One-Class Support Vector Machine Classifier for Reduced Samples

Usually, the One-Class Support Vector Machine (OC-SVM) requires a large dataset for modeling effectively the target class independently to other classes. For finding the OC-SVM model, the available dataset is subdivided into two subsets namely training and validation, which are used for training and validating the optimal parameters. This approach is effective when a large dataset is available. However, when training samples are reduced, parameters of the OC-SVM are difficult to find in absence of the validation subset. Hence, this paper proposes various techniques for selecting the optimal parameters using only a training subset. The experimental evaluation conducted on several real-world benchmarks proves the effective use of the new selection parameter techniques for validating the model of OC-SVM classifiers versus the standard validation techniques

Spatial Coevolution for Generative Adversarial Network Training

Generative Adversarial Networks (GANs) are difficult to train because of pathologies such as mode and discriminator collapse. Similar pathologies have been studied and addressed in competitive evolutionary computation by increased diversity. We study a system, Lipizzaner, that combines spatial coevolution with gradient-based learning to improve the robustness and scalability of GAN training. We study different features of Lipizzaner’s evolutionary computation methodology. Our ablation experiments determine that communication, selection, parameter optimization, and ensemble optimization each, as well as in combination, play critical roles. Lipizzaner succumbs less frequently to critical collapses and, as a side benefit, demonstrates improved performance. In addition, we show a GAN-training feature of Lipizzaner: the ability to train simultaneously with different loss functions in the gradient descent parameter learning framework of each GAN at each cell. We use an image generation problem to show that different loss function combinations result in models with better accuracy and more diversity in comparison to other existing evolutionary GAN models. Finally, Lipizzaner with multiple loss function options promotes the best model diversity while requiring a large grid size for adequate accuracy.

Payoff-dependence learning ability resolves social dilemmas

Understanding the appearance and maintenance of cooperation behavior is one of the most interesting challenges in natural and social sciences. Evolutionary game is a useful tool to study this issue. Here, we consider a basic strategy updating rule: the probability of a player updating its strategy is affected by the learning ability, which is determined by payoffs and an aspiration parameter [Formula: see text]. For positive [Formula: see text], learning ability is directly proportional to player’s own payoff. When [Formula: see text] equals 0, it returns to traditional situation. It is found that increasing the value of [Formula: see text] can promote the cooperation. With the increase of [Formula: see text], the player’s learning ability is continuously enhanced, and the probability of changing strategies is also increased. This paper verifies the influence of the introduced selection parameter [Formula: see text] on the cooperation rate from different aspects. We tested this hypothesis through the Monte Carlo simulation, and demonstrated that introducing [Formula: see text] changed the network of interaction effectively, therefore changing the effect of the adoption of the strategy on the uncertainty of cooperation evolution. This paper analyzed the results of the payoff-dependence learning ability of different players when they imitate the strategies of their opponents, which can effectively promote the evolution of cooperation.

Genetic drift from the out-of-Africa bottleneck leads to biased estimation of genetic architecture and selection

Most complex traits evolved in the ancestors of all modern humans and have been under negative or balancing selection to maintain the distribution of phenotypes observed today. Yet all large studies mapping genomes to complex traits occur in populations that have experienced the Out-of-Africa bottleneck. Does this bottleneck affect the way we characterise complex traits? We demonstrate using the 1000 Genomes dataset and hypothetical complex traits that genetic drift can strongly affect the joint distribution of effect size and SNP frequency, and that the bias can be positive or negative depending on subtle details. Characterisations that rely on this distribution therefore conflate genetic drift and selection. We provide a model to identify the underlying selection parameter in the presence of drift, and demonstrate that a simple sensitivity analysis may be enough to validate existing characterisations. We conclude that biobanks characterising more worldwide diversity would benefit studies of complex traits.

Image processing to identify damage to soybean seeds

ABSTRACT: In Brazil and worldwide, commercialization of soybeans is of great importance to the economy, making their quality considered. The presence of damaged soybean seeds decreases the added value of the product. Businesses need fast and effective techniques to maintain the quality. The present research aimed to identify, through image processing, damage caused to soybean seeds, namely the presence of greenish seeds and wrinkled seeds due to variations of humidity and temperature, where it was possible to identify greenish and wrinkled soybean seeds from images. Results obtained for greenish seeds indicated that the red color scale is the most suitable for selection due to its more significant variation compared to the other color scales. For the separation of wrinkled seeds, it can be stated that it is possible to find a selection parameter with 74.3% accuracy in removing seeds with medium to high degrees of wrinkle damage.

Screening of small molecules using the inhibition of oligomer formation in α-synuclein aggregation as a selection parameter

The aggregation of α-synuclein is a central event in Parkinsons’s disease and related synucleinopathies. Since pharmacologically targeting this process, however, has not yet resulted in approved disease-modifying treatments, there is an unmet need of developing novel methods of drug discovery. In this context, the use of chemical kinetics has recently enabled accurate quantifications of the microscopic steps leading to the proliferation of protein misfolded oligomers. As these species are highly neurotoxic, effective therapeutic strategies may be aimed at reducing their numbers. Here, we exploit this quantitative approach to develop a screening strategy that uses the reactive flux toward α-synuclein oligomers as a selection parameter. Using this approach, we evaluate the efficacy of a library of flavone derivatives, identifying apigenin as a compound that simultaneously delays and reduces the formation of α-synuclein oligomers. These results demonstrate a compound selection strategy based on the inhibition of the formation of α-synuclein oligomers, which may be key in identifying small molecules in drug discovery pipelines for diseases associated with α-synuclein aggregation.

Genetic drift from the out-of-Africa bottleneck leads to biased estimation of genetic architecture and selection

Most complex traits evolved in the ancestors of all modern humans and have been under negative or balancing selection to maintain the distribution of phenotypes observed today. Yet all large studies mapping genomes to complex traits occur in populations that have experienced the Out-of-Africa bottleneck. Does this bottleneck affect the way we characterise complex traits? We demonstrate using the 1000 Genomes dataset and hypothetical complex traits that genetic drift can strongly affect the joint distribution of effect size and SNP frequency. Characterisations that rely on this distribution therefore conflate genetic drift and selection. We provide a model to identify the underlying selection parameter in the presence of drift, and demonstrate that a simple sensitivity analysis may be enough to validate existing characterisations. We conclude that biobanks characterising more worldwide diversity would benefit studies of complex traits.

Introduction of Computer Aided Test Analysis Program for Civil Engineering

The Computer Aided Test Analysis for Civil Engineering (CATACE) is made based on International standards, regulations and specifications. It includes test analysis for aggregates, bitumen, soil, asphalt concrete and its mixtures. It includes the technical requirements for nuclear gauge, selection parameter for surface dressing, blending for prime coat and tack coat. It provides the unit conversion between American to international system. It provided over 40 kinds of test analysis, and it will be increase working efficiency for the Material Testing Engineer who is working oversea project and follows International Standards. The paper introduces the functions and application of the program.