Population interaction network in representative gravitational search algorithms: Logistic distribution leads to worse performance.

In this paper, a novel study on the way inter-individual information interacts in meta-heuristic algorithms (MHAs) is carried out using a scheme known as population interaction networks (PIN). Specifically, three representative MHAs, including the differential evolutionary algorithm (DE), the particle swarm optimization algorithm (PSO), the gravitational search algorithm (GSA), and four classical variations of the gravitational search algorithm, are analyzed in terms of inter-individual information interactions and the differences in the performance of each of the algorithms on IEEE Congress on Evolutionary Computation 2017 benchmark functions. The cumulative distribution function (CDF) of the node degree obtained by the algorithm on the benchmark function is fitted to the seven distribution models by using PIN. The results show that among the seven compared algorithms, the more powerful DE is more skewed towards the Poisson distribution, and the weaker PSO, GSA, and GSA variants are more skewed towards the Logistic distribution. The more deviation from Logistic distribution GSA variants conform, the stronger their performance. From the point of view of the CDF, deviating from the Logistic distribution facilitates the improvement of the GSA. Our findings suggest that the population interaction network is a powerful tool for characterizing and comparing the performance of different MHAs in a more comprehensive and meaningful way.

Integrated transcriptomic and metabolomic analyses revealed the regulatory mechanism of sulfur application in grain yield and protein content in wheat (Triticum aestivum L.).

Sulfur fertilizers play an important role in increasing the yield and improving the dough quality of bread wheat, but their regulatory mechanism remains unclear. In this study, 0 kg·ha-1 (S0) and 60 kg·ha-1 (S60) of sulfur were applied on the anthesis date; subsequently, immature wheat grains at 8, 13, and 18 days post-anthesis (DPA) were subjected to integrated transcriptomic and metabolomic analyses to investigate the changes in the gene/metabolite activity in a typical strong-gluten wheat, Gaoyou2018 (GY2018). Our data show that the S60 treatment could significantly increase the grain yield and grain protein content by 13.2 and 3.6%, respectively. The transcriptomic analysis revealed that 10,694 differentially expressed genes (DEGs) were induced by S60 from 8 to 18 DPA when compared with their corresponding no-sulfur controls, and most DEGs were mainly involved in lipid metabolism and amino acid metabolism pathways. Ninety-seven MYB transcription factors (TFs) were identified as responsive to the S60 treatment; of these, 66 showed significantly differential expression at 13 DPA, and MYB118 might participate in the process of sulfur metabolism by regulating glucosinolate synthesis. In total, 542 significantly enriched differentially expressed (DE) metabolites (DEMs) were identified following the S60 treatment, which mainly included secondary metabolites, carbohydrates, and amino acids. Several metabolites (e.g., glutathione, sucrose, GDP-alpha-D-glucose, and amino acids) exhibited altered abundances following the S60 treatment. The combination of transcriptomic and metabolomic analyses highlighted the important role of amino acid metabolism (especially cysteine, methionine, and glutathione metabolism) and starch and sucrose metabolism pathways after S60 application. Our results provide valuable information enhancing our understanding of the molecular mechanism of the response to sulfur and provide useful clues for grain protein quality formation and yield improvement in bread wheat.

A Ghd7-centered regulatory network provides a mechanistic approximation to optimal heterosis in an elite rice hybrid.

Heterosis refers to the superior performance of hybrids over their parents, which is a general phenomenon occurring in diverse organisms. Many commercial hybrids produce high yield without delayed flowering, which we refer to as optimal heterosis and is desired in hybrid breeding. Here, we attempted to illustrate the genomic basis of optimal heterosis by reinvestigating the single-locus quantitative trait loci and digenic interactions of two traits, the number of spikelets per panicle (SP) and heading date (HD), using recombinant inbred lines and 'immortalized F2 s' derived from the elite rice (Oryza sativa) hybrid Shanyou 63. Our analysis revealed a regulatory network that may provide an approximation to the genetic constitution of the optimal heterosis observed in this hybrid. In this network, Ghd7 works as the core element, and three other genes, Ghd7.1, Hd1, and Hd3a/RFT1, also have major roles. The effects of positive dominance by Ghd7 and Ghd7.1 and negative dominance by Hd1 and Hd3a/RFT1 in the hybrid background contribute the major part to the high SP without delaying HD; numerous epistatic interactions, most of which involve Ghd7, also play important roles collectively. The results expand our understanding of the genic interaction networks underlying hybrid rice breeding programs, which may be very useful in future crop genetic improvement.

Autonomic flexibility reflects learning and associated neuroplasticity in old age.

Effective learning in old age, particularly in those at risk for dementia, is essential for prolonging independent living. Individual variability in learning, however, is remarkable; that is, months of cognitive training to improve learning may be beneficial for some individuals but not others. So far, little is known about which neurophysiological mechanisms account for the observed variability in learning induced by cognitive training in older adults. By combining Lövdén et al.'s (2010, A theoretical framework for the study of adult cognitive plasticity. Psychological Bulletin, 136, 659-676) framework proposing the role of adaptation capacity in neuroplasticity and a neurovisceral integration model of the relationship between autonomic nervous system (ANS) and brain with a novel shapelet analytical approach that allows for accurate and interpretable analysis of time series data, we discovered an acute, ECG-derived ANS segment in response to cognitive training tasks at baseline that predicted learning outcomes from a 6-week cognitive training intervention. The relationship between the ANS segment and learning was robust in both cross-participant and cross-task analyses among a group of older adults with amnestic mild cognitive impairment. Furthermore, the revealed ANS shapelet significantly predicted training-induced neuroplasticity in the dorsal anterior cingulate cortex and select frontal regions during task fMRI. Across outcome measures, individuals were less likely to prospectively benefit from the cognitive training if their ECG data were more similar to this particular ANS segment at baseline. Our findings are among the first empirical evidence to confirm that adaptation capacity, indexed by ANS flexibility, predicts individual differences in learning and associated neuroplasticity beyond individual characteristics (e.g., age, education, neurodegeneration, total training).

A Dendritic Neuron Model with Adaptive Synapses Trained by Differential Evolution Algorithm.

A dendritic neuron model with adaptive synapses (DMASs) based on differential evolution (DE) algorithm training is proposed. According to the signal transmission order, a DNM can be divided into four parts: the synaptic layer, dendritic layer, membrane layer, and somatic cell layer. It can be converted to a logic circuit that is easily implemented on hardware by removing useless synapses and dendrites after training. This logic circuit can be designed to solve complex nonlinear problems using only four basic logical devices: comparators, AND (conjunction), OR (disjunction), and NOT (negation). To obtain a faster and better solution, we adopt the most popular DE for DMAS training. We have chosen five classification datasets from the UCI Machine Learning Repository for an experiment. We analyze and discuss the experimental results in terms of the correct rate, convergence rate, ROC curve, and the cross-validation and then compare the results with a dendritic neuron model trained by the backpropagation algorithm (BP-DNM) and a neural network trained by the backpropagation algorithm (BPNN). The analysis results show that the DE-DMAS shows better performance in all aspects.

A Robust AUC Maximization Framework With Simultaneous Outlier Detection and Feature Selection for Positive-Unlabeled Classification.

The positive-unlabeled (PU) classification is a common scenario in real-world applications such as healthcare, text classification, and bioinformatics, in which we only observe a few samples labeled as "positive" together with a large volume of "unlabeled" samples that may contain both positive and negative samples. Building robust classifiers for the PU problem is very challenging, especially for complex data where the negative samples overwhelm and mislabeled samples or corrupted features exist. To address these three issues, we propose a robust learning framework that unifies area under the curve maximization (a robust metric for biased labels), outlier detection (for excluding wrong labels), and feature selection (for excluding corrupted features). The generalization error bounds are provided for the proposed model that give valuable insight into the theoretical performance of the method and lead to useful practical guidance, e.g., to train a model, we find that the included unlabeled samples are sufficient as long as the sample size is comparable to the number of positive samples in the training process. Empirical comparisons and two real-world applications on surgical site infection (SSI) and EEG seizure detection are also conducted to show the effectiveness of the proposed model.

Data-dependent hashing based on p-stable distribution.

The p-stable distribution is traditionally used for data-independent hashing. In this paper, we describe how to perform data-dependent hashing based on p-stable distribution. We commence by formulating the Euclidean distance preserving property in terms of variance estimation. Based on this property, we develop a projection method, which maps the original data to arbitrary dimensional vectors. Each projection vector is a linear combination of multiple random vectors subject to p-stable distribution, in which the weights for the linear combination are learned based on the training data. An orthogonal matrix is then learned data-dependently for minimizing the thresholding error in quantization. Combining the projection method and orthogonal matrix, we develop an unsupervised hashing scheme, which preserves the Euclidean distance. Compared with data-independent hashing methods, our method takes the data distribution into consideration and gives more accurate hashing results with compact hash codes. Different from many data-dependent hashing methods, our method accommodates multiple hash tables and is not restricted by the number of hash functions. To extend our method to a supervised scenario, we incorporate a supervised label propagation scheme into the proposed projection method. This results in a supervised hashing scheme, which preserves semantic similarity of data. Experimental results show that our methods have outperformed several state-of-the-art hashing approaches in both effectiveness and efficiency.