BACKGROUND: One possible approach how to economically facilitate gene expression profiling is to use the L1000 platform which measures the expression of ∼1,000 landmark genes and uses a computational method to infer the expression of another ∼10,000 genes. One such method for the gene expression inference is a D-GEX which employs neural networks. RESULTS: We propose two novel D-GEX architectures that significantly improve the quality of the inference by increasing the capacity of a network without any increase in the number of trained parameters. The architectures partition the network into individual towers. Our best proposed architecture - a checkerboard architecture with a skip connection and five towers - together with minor changes in the training protocol improves the average mean absolute error of the inference from 0.134 to 0.128. CONCLUSIONS: Our proposed approach increases the gene expression inference accuracy without increasing the number of weights of the model and thus without increasing the memory footprint of the model that is limiting its usage.

• Klíčová slova
• Checkerboard architecture, Gene expression, Neural network, Tower architecture,
• MeSH
• algoritmy MeSH
• exprese genu MeSH
• genové regulační sítě MeSH
• neuronové sítě * MeSH
• stanovení celkové genové exprese * MeSH
• Publikační typ
• časopisecké články MeSH

The technologies of the Internet of Things (IoT) have an increasing influence on our daily lives. The expansion of the IoT is associated with the growing number of IoT devices that are connected to the Internet. As the number of connected devices grows, the demand for speed and data volume is also greater. While most IoT network technologies use cloud computing, this solution becomes inefficient for some use-cases. For example, suppose that a company that uses an IoT network with several sensors to collect data within a production hall. The company may require sharing only selected data to the public cloud and responding faster to specific events. In the case of a large amount of data, the off-loading techniques can be utilized to reach higher efficiency. Meeting these requirements is difficult or impossible for solutions adopting cloud computing. The fog computing paradigm addresses these cases by providing data processing closer to end devices. This paper proposes three possible network architectures that adopt fog computing for LoRaWAN because LoRaWAN is already deployed in many locations and offers long-distance communication with low-power consumption. The architecture proposals are further compared in simulations to select the optimal form in terms of total service time. The resulting optimal communication architecture could be deployed to the existing LoRaWAN with minimal cost and effort of the network operator.

• Klíčová slova
• LoRaWAN, cloud computing, fog computing, internet of things, network architecture, simulation,
• Publikační typ
• časopisecké články MeSH

Objective.The current practices of designing neural networks rely heavily on subjective judgment and heuristic steps, often dictated by the level of expertise possessed by architecture designers. To alleviate these challenges and streamline the design process, we propose an automatic method, a novel approach to enhance the optimization of neural network architectures for processing intracranial electroencephalogram (iEEG) data.Approach.We present a genetic algorithm, which optimizes neural network architecture and signal pre-processing parameters for iEEG classification.Main results.Our method improved the macroF1 score of the state-of-the-art model in two independent datasets, from St. Anne's University Hospital (Brno, Czech Republic) and Mayo Clinic (Rochester, MN, USA), from 0.9076 to 0.9673 and from 0.9222 to 0.9400 respectively.Significance.By incorporating principles of evolutionary optimization, our approach reduces the reliance on human intuition and empirical guesswork in architecture design, thus promoting more efficient and effective neural network models. The proposed method achieved significantly improved results when compared to the state-of-the-art benchmark model (McNemar's test,p≪ 0.01). The results indicate that neural network architectures designed through machine-based optimization outperform those crafted using the subjective heuristic approach of a human expert. Furthermore, we show that well-designed data preprocessing significantly affects the models' performance.

• Klíčová slova
• deep learning, genetic algorithms, intracranial EEG, neural network, optimization,
• MeSH
• elektroencefalografie metody   MeSH
• elektrokortikografie * MeSH
• lidé MeSH
• neuronové sítě * MeSH
• počítačové zpracování signálu MeSH
• Check Tag
• lidé MeSH
• Publikační typ
• časopisecké články MeSH
• práce podpořená grantem MeSH

Cognitive flexibility is a major requirement for successful behavior. nNeural oscillations in the alpha frequency band were repeatedly associated with cognitive flexibility in task-switching paradigms. Alpha frequencies are modulated by working memory load and are used to process information during task switching, however we do not know how this oscillatory network communication is modulated. In order to understand the mechanisms that drive cognitive flexibility, ERPs, oscillatory power and how the communication within these networks is organized are of importance. The EEG data show that during phases reflecting preparatory processes to pre-activate task sets, alpha oscillatory power but not the small world properties of the alpha network architecture was modulated. During the switching only the N2 ERP component showed clear modulations. After the response, alpha oscillatory power reinstates and therefore seems to be important to deactivate or maintain the previous task set. For these reactive control processes the network architecture in terms of small-world properties is modulated. Effects of memory load on small-world aspects were seen in repetition trials, where small-world properties were higher when memory processes were relevant. These results suggest that the alpha oscillatory network becomes more small-world-like when reactive control processes during task switching are less complex.

• MeSH
• dospělí MeSH
• elektroencefalografie * MeSH
• evokované potenciály MeSH
• kognice fyziologie   MeSH
• lidé středního věku MeSH
• lidé MeSH
• mapování mozku MeSH
• mladiství MeSH
• mladý dospělý MeSH
• paměť fyziologie   MeSH
• reakční čas MeSH
• zdraví dobrovolníci pro lékařské studie MeSH
• Check Tag
• dospělí MeSH
• lidé středního věku MeSH
• lidé MeSH
• mladiství MeSH
• mladý dospělý MeSH
• mužské pohlaví MeSH
• ženské pohlaví MeSH
• Publikační typ
• časopisecké články MeSH
• práce podpořená grantem MeSH

Verification of the correct functionality of multi-vehicle systems in high-fidelity scenarios is required before any deployment of such a complex system, e.g., in missions of remote sensing or in mobile sensor networks. Mixed-reality simulations where both virtual and physical entities can coexist and interact have been shown to be beneficial for development, testing, and verification of such systems. This paper deals with the problems of designing a certain communication subsystem for such highly desirable realistic simulations. Requirements of this communication subsystem, including proper addressing, transparent routing, visibility modeling, or message management, are specified prior to designing an appropriate solution. Then, a suitable architecture of this communication subsystem is proposed together with solutions to the challenges that arise when simultaneous virtual and physical message transmissions occur. The proposed architecture can be utilized as a high-fidelity network simulator for vehicular systems with implicit mobility models that are given by real trajectories of the vehicles. The architecture has been utilized within multiple projects dealing with the development and practical deployment of multi-UAV systems, which support the architecture's viability and advantages. The provided experimental results show the achieved similarity of the communication characteristics of the fully deployed hardware setup to the setup utilizing the proposed mixed-reality architecture.

• Klíčová slova
• communication architecture, middleware, mixed-reality simulations, testbeds, unmanned systems,
• Publikační typ
• časopisecké články MeSH

BACKGROUND: The recent big data revolution in Genomics, coupled with the emergence of Deep Learning as a set of powerful machine learning methods, has shifted the standard practices of machine learning for Genomics. Even though Deep Learning methods such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) are becoming widespread in Genomics, developing and training such models is outside the ability of most researchers in the field. RESULTS: Here we present ENNGene-Easy Neural Network model building tool for Genomics. This tool simplifies training of custom CNN or hybrid CNN-RNN models on genomic data via an easy-to-use Graphical User Interface. ENNGene allows multiple input branches, including sequence, evolutionary conservation, and secondary structure, and performs all the necessary preprocessing steps, allowing simple input such as genomic coordinates. The network architecture is selected and fully customized by the user, from the number and types of the layers to each layer's precise set-up. ENNGene then deals with all steps of training and evaluation of the model, exporting valuable metrics such as multi-class ROC and precision-recall curve plots or TensorBoard log files. To facilitate interpretation of the predicted results, we deploy Integrated Gradients, providing the user with a graphical representation of an attribution level of each input position. To showcase the usage of ENNGene, we train multiple models on the RBP24 dataset, quickly reaching the state of the art while improving the performance on more than half of the proteins by including the evolutionary conservation score and tuning the network per protein. CONCLUSIONS: As the role of DL in big data analysis in the near future is indisputable, it is important to make it available for a broader range of researchers. We believe that an easy-to-use tool such as ENNGene can allow Genomics researchers without a background in Computational Sciences to harness the power of DL to gain better insights into and extract important information from the large amounts of data available in the field.

• Klíčová slova
• Convolutional Neural Network, Deep Learning, Evolutionary Conservation Score, GUI, RNA Secondary Structure, Recurrent Neural Network,
• MeSH
• genomika MeSH
• neuronové sítě * MeSH
• sekundární struktura proteinů MeSH
• strojové učení * MeSH
• Publikační typ
• časopisecké články MeSH

Spiking Neural Networks (SNNs) have been considered a potential competitor to Artificial Neural Networks (ANNs) due to their high biological plausibility and energy efficiency. However, the architecture design of SNN has not been well studied. Previous studies either use ANN architectures or directly search for SNN architectures under a highly constrained search space. In this paper, we aim to introduce much more complex connection topologies to SNNs to further exploit the potential of SNN architectures. To this end, we propose the topology-aware search space, which is the first search space that enables a more diverse and flexible design for both the spatial and temporal topology of the SNN architecture. Then, to efficiently obtain architecture from our search space, we propose the spatio-temporal topology sampling (STTS) algorithm. By leveraging the benefits of random sampling, STTS can yield powerful architecture without the need for an exhaustive search process, making it significantly more efficient than alternative search strategies. Extensive experiments on CIFAR-10, CIFAR-100, and ImageNet demonstrate the effectiveness of our method. Notably, we obtain 70.79% top-1 accuracy on ImageNet with only 4 time steps, 1.79% higher than the second best model. Our code is available under https://github.com/stiger1000/Random-Sampling-SNN.

• Klíčová slova
• Neural architecture search, Spiking neural networks,
• MeSH
• algoritmy * MeSH
• neuronové sítě * MeSH
• Publikační typ
• časopisecké články MeSH

A new approach with artificial neural network (ANN) was applied to numerical taxonomy of bacteria using the oxalate as carbon and energy source. For this aim the characters effective in differentiating separate groups were selected from morphological, physiological and biochemical test results. Fourteen aerobic, Gram-negative, oxalate-utilizing isolates and four oxalate-utilizing reference strains (Ralstonia eutropha DSM 428, Methylobacterium extorquens DSM 1337T, Ralstonia oxalatica DSM 1105T, Oxalicibacterium flavum DSM 15506T) were included in the study. ANN program used here was developed in Borland C++ language. Iterations were performed on an IBM compatible PC computer. ANN architecture having feed-forward backpropagation algorithm was used for teaching generalized delta rule. The results show that ANN can have a large potential in solving the taxonomic problems of oxalate-utilizing bacteria.

• MeSH
• gramnegativní aerobní bakterie klasifikace   metabolismus   MeSH
• neuronové sítě * MeSH
• oxaláty metabolismus   MeSH
• Publikační typ
• časopisecké články MeSH
• Názvy látek
• oxaláty MeSH

Diet, stress, genetics, and a sedentary lifestyle may all contribute to heart disease rates. Although recent studies propose comprehensive automated diagnostic systems, these systems tend to focus on one aspect, such as feature selection, prioritization, or predictive accuracy. A more complete approach that considers all of these factors can improve the efficiency of a cardiac prediction system. This study uses an appropriate strategy to overcome potential network design problems, design challenges, overfitting, and lack of robustness that can interfere with system performance. The research introduces an ideally designed deep trust network called ID-DTN to improve system performance. The Ruzzo-Tompa method is used to eliminate noncontributory features. The Seagull Optimization Algorithm (SOA) is introduced to optimize the trust depth network to achieve optimal network design. The study scrutinizes the deep trust network (ID-DTN) and the restricted Boltzmann machine (RBM) and sheds light on the system's operation. This proposal can optimize both network architecture and feature selection, which is the main novelty. The proposed method is analyzed using the below-mentioned metrics: Matthew's correlation coefficient, F1 score, accuracy, sensitivity, specificity, and accuracy. ID-DTN performs well compared to other state-of-the-art methods. The validation results confirm that the proposed method improves the prediction accuracy to 97.11% and provides reliable recommendations for patients with cardiovascular disease.

• Klíčová slova
• Artificial Intelligence, Boltzmann machine, Deep learning, Heart disease prediction, Ruzzo-Tompa, Seagull optimization,
• MeSH
• algoritmy * MeSH
• lidé MeSH
• nemoci srdce * diagnóza   MeSH
• neuronové sítě MeSH
• Check Tag
• lidé MeSH
• Publikační typ
• časopisecké články MeSH

Metamaterials present the possibility of artificially generating advanced functionalities through engineering of their internal structure. Artificial spin networks, in which a large number of nanoscale magnetic elements are coupled together, are promising metamaterial candidates that enable the control of collective magnetic behavior through tuning of the local interaction between elements. In this work, the motion of magnetic domain-walls in an artificial spin network leads to a tunable stochastic response of the metamaterial, which can be tailored through an external magnetic field and local lattice modifications. This type of tunable stochastic network produces a controllable random response exploiting intrinsic stochasticity within magnetic domain-wall motion at the nanoscale. An iconic demonstration used to illustrate the control of randomness is the Galton board. In this system, multiple balls fall into an array of pegs to generate a bell-shaped curve that can be modified via the array spacing or the tilt of the board. A nanoscale recreation of this experiment using an artificial spin network is employed to demonstrate tunable stochasticity. This type of tunable stochastic network opens new paths toward post-Von Neumann computing architectures such as Bayesian sensing or random neural networks, in which stochasticity is harnessed to efficiently perform complex computational tasks.

• Klíčová slova
• Galton board, artificial spin network, computing, magnetic domain-wall, metamaterial, tunable stochasticity,
• Publikační typ
• časopisecké články MeSH