All issues

In today’s technology-driven world, lite devices like Internet of Things (IoT) devices and microcontrollers (MCUs) are becoming increasingly common. These devices are more energyefficient and affordable, often with reduced features compared to the standard versions such as very limited memory and processing power for typical machine learning models. However, modern machine learning models can have millions of parameters, resulting in a large memory footprint. This complexity not only makes it difficult to deploy these large models on resource constrained devices but also increases the risk of latency and inefficiency in processing, which is crucial in some cases where real-time responses are required such as autonomous driving and medical diagnostics. In recent years, neural networks have seen significant advancements in model optimization techniques that help deployment and inference on these small devices. This narrative review offers a thorough examination of the progression and latest developments in neural network optimization, focusing on key areas such as quantization, pruning, knowledge distillation, and neural architecture search. It examines how these algorithmic solutions have progressed and how new approaches have improved upon the existing techniques making neural networks more efficient. This review is designed for machine learning researchers, practitioners, and engineers who may be unfamiliar with these methods but wish to explore the available techniques. It highlights ongoing research in optimizing networks for achieving better performance, lowering energy consumption, and enabling faster training times, all of which play an important role in the continued scalability of neural networks. Additionally, it identifies gaps in current research and provides a foundation for future studies, aiming to enhance the applicability and effectiveness of existing optimization strategies.

This paper investigates the effectiveness of Retrieval-Augmented Generation (RAG) combined with various Large Language Models (LLMs) for document retrieval and information access in corporate information systems. We survey typical use-cases of LLMs in enterprise environments, outline the RAG architecture, and discuss the major challenges that arise when integrating LLMs into a RAG pipeline. A system architecture is proposed that couples a text-vector encoder with an LLM. The encoder builds a vector database that indexes a library of corporate documents. For every user query, relevant contextual fragments are retrieved from this library via the FAISS engine and appended to the prompt given to the LLM. The LLM then generates an answer grounded in the supplied context. The overall structure and workflow of the proposed RAG solution are described in detail. To justify the choice of the generative component, we benchmark a set of widely used LLMs — ChatGPT, GigaChat, YandexGPT, Llama, Mistral, Qwen, and others — when employed as the answer-generation module. Using an expert-annotated test set of queries, we evaluate the accuracy, completeness, linguistic quality, and conciseness of the responses. Model-specific characteristics and average response latencies are analysed; the study highlights the significant influence of available GPU memory on the throughput of local LLM deployments. An overall ranking of the models is derived from an aggregated quality metric. The results confirm that the proposed RAG architecture provides efficient document retrieval and information delivery in corporate environments. Future research directions include richer context augmentation techniques and a transition toward agent-based LLM architectures. The paper concludes with practical recommendations on selecting an optimal RAG–LLM configuration to ensure fast and precise access to enterprise knowledge assets.

We schedule the computer technology for present the engineer psychology tests which reveal probationer men which may hasten the logic task solution by simultaneous execution several standard logic operations. These tests based on the theory of two logic task kinds: in first kind the parallel logic is effectively, and in second kind it is not effectively. The realize experiment confirms the capability parallel logic for impotent part of people. The vital speedup execution of logic operations is very uncommon in simultaneous logic. The efficacy of methodic is confirmed.

The aim of the present study is to show how engineering approaches and ideas work in clinical restorative dentistry, in particular, how they affect the restoration design and durability of restored endodontically treated teeth. For these purposes a 3D-computational model of a first incisor including the elements of hard tooth tissues, periodontal ligament, surrounding bone structures and restoration itself has been constructed and numerically simulated for a variety of restoration designs under normal chewing loadings. It has been researched the effect of different adhesive technologies in root canal on the functional characteristics of a restored tooth. The 3D model designed could be applied for preclinical diagnostics to determine the areas of possible fractures of a restored tooth and prognosticate its longevity.

The paper presents a physico-mathematical model of the perturbed region formed in the lower D-layer of the ionosphere under the action of directed radio emission flux from a terrestrial stand of the megahertz frequency range, obtained as a result of comprehensive theoretical studies. The model is based on the consideration of a wide range of kinetic processes taking into account their nonequilibrium and in the two-temperature approximation for describing the transformation of the radio beam energy absorbed by electrons. The initial data on radio emission achieved by the most powerful radio-heating stands are taken in the paper. Their basic characteristics and principles of functioning, and features of the altitude distribution of the absorbed electromagnetic energy of the radio beam are briefly described. The paper presents the decisive role of the D-layer of the ionosphere in the absorption of the energy of the radio beam. On the basis of theoretical analysis, analytical expressions are obtained for the contribution of various inelastic processes to the distribution of the absorbed energy, which makes it possible to correctly describe the contribution of each of the processes considered. The work considers more than 60 components. The change of the component concentration describe about 160 reactions. All the reactions are divided into five groups according to their physical content: ionization-chemical block, excitation block of metastable electronic states, cluster block, excitation block of vibrational states and block of impurities. Blocks are interrelated and can be calculated both jointly and separately. The paper presents the behavior of the parameters of the perturbed region in daytime and nighttime conditions is significantly different at the same radio flux density: under day conditions, the maximum electron concentration and temperature are at an altitude of ~45–55 km; in night ~80 km, with the temperature of heavy particles rapidly increasing, which leads to the occurrence of a gas-dynamic flow. Therefore, a special numerical algorithm are developed to solve two basic problems: kinetic and gas dynamic. Based on the altitude and temporal behavior of concentrations and temperatures, the algorithm makes it possible to determine the ionization and emission of the ionosphere in the visible and infrared spectral range, which makes it possible to evaluate the influence of the perturbed region on radio engineering and optoelectronic devices used in space technology.

We describe an engineering-psychological experiment that continues the study of ways to adapt a person to the increasing complexity of logical problems by presenting a series of problems of increasing complexity, which is determined by the volume of initial data. Tasks require calculations in an associative or non-associative system of operations. By the nature of the change in the time of solving the problem, depending on the number of necessary operations, we can conclude that a purely sequential method of solving problems or connecting additional brain resources to the solution in parallel mode. In a previously published experimental work, a person in the process of solving an associative problem recognized color images with meaningful images. In the new study, a similar problem is solved for abstract monochrome geometric shapes. Analysis of the result showed that for the second case, the probability of the subject switching to a parallel method of processing visual information is significantly reduced. The research method is based on presenting a person with two types of tasks. One type of problem contains associative calculations and allows a parallel solution algorithm. Another type of problem is the control one, which contains problems in which calculations are not associative and parallel algorithms are ineffective. The task of recognizing and searching for a given object is associative. A parallel strategy significantly speeds up the solution with relatively small additional resources. As a control series of problems (to separate parallel work from the acceleration of a sequential algorithm), we use, as in the previous experiment, a non-associative comparison problem in cyclic arithmetic, presented in the visual form of the game “rock, paper, scissors”. In this problem, the parallel algorithm requires a large number of processors with a small efficiency coefficient. Therefore, the transition of a person to a parallel algorithm for solving this problem is almost impossible, and the acceleration of processing input information is possible only by increasing the speed. Comparing the dependence of the solution time on the volume of source data for two types of problems allows us to identify four types of strategies for adapting to the increasing complexity of the problem: uniform sequential, accelerated sequential, parallel computing (where possible), or undefined (for this method) strategy. The Reducing of the number of subjects, who switch to a parallel strategy when encoding input information with formal images, shows the effectiveness of codes that cause subject associations. They increase the speed of human perception and processing of information. The article contains a preliminary mathematical model that explains this phenomenon. It is based on the appearance of a second set of initial data, which occurs in a person as a result of recognizing the depicted objects.

The temporal variability of exergy of short-wave and long-wave radiation and its relationships with sensible heat, water vapor (H₂O) and carbon dioxide (CO₂) fluxes on a recently clear-cut area in a mixed coniferous and small-leaved forest in the Tver region is discussed. On the basis of the analysis of radiation and exergy efficiency coefficients suggested by Yu.M. Svirezhev it was shown that during the first eight months after clearcutting the forest ecosystem functions as a "heat engine" i.e. the processes of energy dissipation dominated over processes of biomass production. To validate the findings the statistical analysis of temporary variability of meteorological parameters, as well as, daily fluxes of sensible heat, H₂O and CO₂ was provided using the trigonometrical polynomials. The statistical models that are linearly depended on an exergy of short-wave and long-wave radiation were obtained for mean daily values of CO₂ fluxes, gross primary production of regenerated vegetation and sensible heat fluxes. The analysis of these dependences is also confirmed the results obtained from processing the radiation and exergy efficiency coefficients. The splitting the time series into separate time intervals, e.g. “spring–summer” and “summer–autumn”, allowed revealing that the statistically significant relationships between atmospheric fluxes and exergy were amplified in summer months as the clear-cut area was overgrown by grassy and young woody vegetation. The analysis of linear relationships between time-series of latent heat fluxes and exergy showed their statistical insignificance. The linear relationships between latent heat fluxes and temperature were in turn statistically significant. The air temperature was a key factor improving the accuracy of the models, whereas effect of exergy was insignificant. The results indicated that at the time of active vegetation regeneration within the clear-cut area the seasonal variability of surface evaporation is mainly governed by temperature variation.

The work continues research on the ability of a person to improve the productivity of information processing, using parallel work or improving the performance of analyzers. A person receives a series of tasks, the solution of which requires the processing of a certain amount of information. The time and the validity of the decision are recorded. The dependence of the average solution time on the amount of information in the problem is determined by correctly solved problems. In accordance with the proposed method, the problems contain calculations of expressions in two algebras, one of which is associative and the other is nonassociative. To facilitate the work of the subjects in the experiment were used figurative graphic images of elements of algebra. Non-associative calculations were implemented in the form of the game “rock-paper-scissors”. It was necessary to determine the winning symbol in the long line of these figures, considering that they appear sequentially from left to right and play with the previous winner symbol. Associative calculations were based on the recognition of drawings from a finite set of simple images. It was necessary to determine which figure from this set in the line is not enough, or to state that all the pictures are present. In each problem there was no more than one picture. Computation in associative algebra allows the parallel counting, and in the absence of associativity only sequential computations are possible. Therefore, the analysis of the time for solving a series of problems reveals a consistent uniform, sequential accelerated and parallel computing strategy. In the experiments it was found that all subjects used a uniform sequential strategy to solve non-associative problems. For the associative task, all subjects used parallel computing, and some have used parallel computing acceleration of the growth of complexity of the task. A small part of the subjects with a high complexity, judging by the evolution of the solution time, supplemented the parallel account with a sequential stage of calculations (possibly to control the solution). We develop a special method for assessing the rate of processing of input information by a person. It allowed us to estimate the level of parallelism of the calculation in the associative task. Parallelism of level from two to three was registered. The characteristic speed of information processing in the sequential case (about one and a half characters per second) is twice less than the typical speed of human image recognition. Apparently the difference in processing time actually spent on the calculation process. For an associative problem in the case of a minimum amount of information, the solution time is near to the non-associativity case or less than twice. This is probably due to the fact that for a small number of characters recognition almost exhausts the calculations for the used non-associative problem.

The repressilator is the first genetic regulatory network in synthetic biology, which was artificially constructed in 2000. It is a closed network of three genetic elements $lacI$, $\lambda cI$ and $tetR$, which have a natural origin, but are not found in nature in such a combination. The promoter of each of the three genes controls the next cistron via the negative feedback, suppressing the expression of the neighboring gene. In our previous paper [Bratsun et al., 2018], we proposed a mathematical model of a delayed repressillator and studied its properties within the framework of a deterministic description. We assume that delay can be both natural, i.e. arises during the transcription / translation of genes due to the multistage nature of these processes, and artificial, i.e. specially to be introduced into the work of the regulatory network using gene engineering technologies. In this work, we apply the stochastic description of dynamic processes in a delayed repressilator, which is an important addition to deterministic analysis due to the small number of molecules involved in gene regulation. The stochastic study is carried out numerically using the Gillespie algorithm, which is modified for time delay systems. We present the description of the algorithm, its software implementation, and the results of benchmark simulations for a onegene delayed autorepressor. When studying the behavior of a repressilator, we show that a stochastic description in a number of cases gives new information about the behavior of a system, which does not reduce to deterministic dynamics even when averaged over a large number of realizations. We show that in the subcritical range of parameters, where deterministic analysis predicts the absolute stability of the system, quasi-regular oscillations may be excited due to the nonlinear interaction of noise and delay. Earlier, we have discovered within the framework of the deterministic description, that there exists a long-lived transient regime, which is represented in the phase space by a slow manifold. This mode reflects the process of long-term synchronization of protein pulsations in the work of the repressilator genes. In this work, we show that the transition to the cooperative mode of gene operation occurs a two order of magnitude faster, when the effect of the intrinsic noise is taken into account. We have obtained the probability distribution of moment when the phase trajectory leaves the slow manifold and have determined the most probable time for such a transition. The influence of the intrinsic noise of chemical reactions on the dynamic properties of the repressilator is discussed.