Abstract Source code vulnerability detection is a crucial aspect of software security development, and the current use of Large Language Models (LLMs) accelerates not only software development but also the generation and propagation of code vulnerabilities. Traditional code vulnerability detection techniques have limited detection efficiency and accuracy. Deep learning techniques have recently gained distinct advantages in multidimensional feature extraction and large-scale data processing, and their application in code vulnerability detection is evolving from simple classification to multimodal approaches. This paper primarily systematizes and summarises deep learning-based source code vulnerability detection, as well as analyzes and anticipates current challenges and future research directions in this area. The distinction between this review and the preceding reviews: This study investigates the literature of the last four years; Not only does it contain datasets, but also includes model-related research and an analysis of multiple different application scenarios. It’s more current and comprehensive than most previous reviews of this type.
El presente estudio se justifica por la necesidad de mejorar las propiedades del mortero de cemento portland frente a la humedad, uno de los principales problemas en edificaciones ubicadas en zonas húmedas, que afecta tanto la durabilidad como la habitabilidad de las viviendas. Se propuso evaluar el impacto de la adición de resina de colofonia sobre las propiedades mecánico-físicas del mortero utilizado en muros de albañilería confinada. El objetivo principal fue determinar si esta adición mejora la impermeabilidad y adherencia sin comprometer significativamente la resistencia estructural. Para ello, se llevaron a cabo ensayos conforme a normas técnicas peruanas y chilenas, que incluyeron pruebas de compresión en prismas, compresión diagonal en muretes, resistencia a la flexión, absorción de agua y adherencia a cizalle, utilizando morteros con porcentajes de colofonia de 0.2%, 0.4% y 0.5% en diferentes dosificaciones. Los resultados evidenciaron que la incorporación de colofonia reduce la resistencia a la compresión y flexión, así como la rigidez del mortero; sin embargo, mejora notablemente la impermeabilidad y la adherencia, especialmente en dosificaciones con 0.2% y 0.4% de resina. Se concluyó que, aunque el uso de colofonia disminuye ciertas propiedades mecánicas, sus beneficios físicos como la reducción de la absorción capilar y el aumento de la adherencia la convierten en una alternativa viable en viviendas de hasta dos pisos, en contextos donde el control de humedad es prioritario sobre la resistencia estructural.
Se diseñó una metodología computacional para evaluar la motivación y la retención del conocimiento en educación superior, utilizando técnicas de análisis estadístico avanzado y modelos de gamificación. El objetivo fue desarrollar un sistema que permita medir cuantitativamente el impacto de estrategias gamificadas en la motivación y la memoria de los estudiantes. La intervención se llevó a cabo en un estudio de caso con 60 estudiantes de educación superior, durante un período de 16 semanas, implementando actividades gamificadas con desafíos, recompensas y retroalimentación en tiempo real. El modelo computacional fue validado usando análisis de varianza (ANOVA) y análisis correlacional, evaluando los cambios en los niveles de motivación, participación y retención del conocimiento antes y después de la intervención. Los resultados mostraron un incremento estadísticamente significativo en la motivación y en las puntuaciones de los conocimientos retenidos, corroborando la efectividad de las estrategias gamificadas aplicadas. Las conclusiones indican que la metodología es una herramienta efectiva para potenciar el aprendizaje activo y mejorar la retención del conocimiento, evidenciando beneficios claros en el rendimiento académico. Sin embargo, se recomienda fortalecer la capacitación docente y mejorar las capacidades tecnológicas en futuras implementaciones para optimizar aún más sus resultados y sostenibilidad. En general, el estudio confirma que la gamificación, cuando se aplica adecuadamente, puede transformar los procesos educativos en entornos superiores.
Thales B. Paiva, Gabrielle De Micheli, Syed Mahbub Hafiz
et al.
Amortized bootstrapping techniques have been proposed for FHEW/TFHE to efficiently refresh multiple ciphertexts simultaneously within a polynomial modulus. Although recent proposals have very efficient asymptotic complexity, reducing the amortized cost essentially to Õ(1) FHE multiplications, the practicality of such algorithms still suffers from substantial overhead and high decryption failure rates (DFR). In this study, we improve upon one of the state-of-the-art amortized bootstrapping algorithms (Guimarães et al., ASIACRYPT 2023) for FHEW/TFHE-like schemes by introducing an alternative algorithmic strategy. Specifically, we combine Guimarães et al.’s strategy based on a two-part NTT with an incomplete Number Theoretic Transform (NTT) algorithm. The resulting construction is such that the multiplication of higher-degree polynomials that would usually create a bottleneck in an incomplete NTT setting actually comes for free. As a result, we demonstrate a 2.12x speedup compared to the algorithm of Guimarães et al. and a 1.12x improvement over the state-of-the-art (sequential) TFHE-rs while achieving a DFR close to 2−32 for 7-bit messages, although the DFR is higher for 8-bit messages. We also explore trade-offs between execution time and DFR, identifying parameter sets that improve the execution time of Guimarães et al. by 1.41x, while simultaneously reducing the DFR by a factor of 2−22 for 8-bit messages.
Computer engineering. Computer hardware, Information technology
Early detection of skin cancer ensures the survival of many cases. There are still challenges in segmenting dermoscopic skin lesion images. Artifacts in the lesion images, such as various dirt, hairs, low contrast, and unclear boundaries, are challenges that affect segmentation accuracy. Convolutional neural networks have brought success in skin lesion segmentation. U‐shaped and V‐shaped deep learning‐based segmentation architectures learn boundary information in the first layers. However, this information becomes weaker in the following layers. Herein, the Edge‐aTtention module is added to the V‐Net architecture to move edge information to the last layer, and the spatial and channel squeeze‐excitation module is added to emphasize high‐level features by recalibrating the channel information to learn lesion boundaries better. The scSEETV‐Net is supported by fusing the binary cross‐entropy, which calculates the loss on a pixel‐based, and the focal Twersky loss function, which has significant success in class imbalances. The proposed architecture achieves 0.9212 Jaccard and 0.9552 Dice in the ISIC2016 dataset, 0.8273 Jaccard and 0.8949 Dice in the ISIC2017 dataset, and 0.8070 Jaccard and 0.8831 Dice in the ISIC2018 dataset. Comparative analyses show that the proposed methodology outperforms the state‐of‐the‐art techniques in the literature.
Computer engineering. Computer hardware, Control engineering systems. Automatic machinery (General)
The existing research on roundabout traffic optimization is mostly based on lossless communication and combines fundamental data, such as vehicle speed, to design collaborative strategies.This approach overlooks data such as the external environment that affects traffic strategies and cannot meet the needs of practical applications.To solve the problem of practical communication constraints, a collaborative lane-change strategy for vehicles around the roundabout based on vehicle state prediction combined with user portraits is proposed.This strategy takes into account the characteristics of vehicle-road-environment collaboration of intelligent connected vehicle.A vehicle prediction method based on spatiotemporal features AP-LSTM, is designed to capture key spatiotemporal features of vehicles to achieve small sample trajectory prediction, effectively improving the accuracy and real-time performance of small sample vehicle trajectory prediction.A predictive mechanism-based vehicle collaboration algorithm PMC is also designed to compensate for the missing vehicle status information in vehicle collaboration decision-making under real-time communication constraints.The future state of vehicles is predicted based on historical data. On this basis, collaborative vehicle control at roundabouts is performed in combination with the Stackelberg game.Simulation experiments are conducted on the SUMO platform.The results show that the proposed AP-LSTM prediction algorithm has a lower Root Mean Square Error(RMSE) value than the Long Short-Term Memory (LSTM) algorithm.At the same time, compared to the SUMO algorithm, the proposed PMC collaborative algorithm has a 51.7% reduction in acceleration standard deviation and an average speed increase of 3.0%, effectively improving the traffic efficiency and driving stability of roundabout traffic.
LI Junhuai, CHEN Miaomiao, WANG Huaijun, CUI Ying'an, ZHANG Aihua
Named Entity Recognition(NER) is an important basis for upper-level natural language processing tasks such as knowledge graph construction, search engines, and recommendation systems.Chinese NER labels and classifies proper nouns or specific named entities in a text sequence.Aiming at the problem that the existing Chinese NER methods cannot effectively extract long-distance semantic information and solve the problem of polysemy, this study proposes a Chinese NER method based on ALBERT pre-training language model, Bidirectional Gated Recurrent Unit(BGRU) and Conditional Random Field(CRF), called ALBERT-BGRU-CRF model.First, the ALBERT pre-trained language model performs word embedding on the input text to obtain dynamic word vectors, which can effectively solve the polysemy problem.Second, BGRU extracts contextual semantic features to further understand semantics and obtain semantic features between long-distance words.Finally, the concatenated vector is input to the CRF layer and decoded using the Viterbi algorithm to reduce the probability of wrongly labelling the output.Then, the entity annotation information is obtained, and the Chinese NER task is completed.The experimental results show that the Chinese NER accuracy and recall rate of the ALBERT-BGRU-CRF model on the MSRA corpus reach 95.16% and 94.58%, respectively.Simultaneously, compared with the fragment neural network model and the CNN-BiLSTM-CRF model, the F1 value of the ALBERT-BGRU-CRF model has increased by 4.43 and 3.78 percentage points.
Yevhenii O. Modlo, S. Semerikov, S. Bondarevskyi
et al.
An analysis of the experience of professional training bachelors of electromechanics in Ukraine and abroad made it possible to determine that one of the leading trends in its modernization is the synergistic integration of various engineering branches (mechanical, electrical, electronic engineering and automation) in mechatronics for the purpose of design, manufacture, operation and maintenance electromechanical equipment. Teaching mechatronics provides for the meaningful integration of various disciplines of professional and practical training bachelors of electromechanics based on the concept of modeling and technological integration of various organizational forms and teaching methods based on the concept of mobility. Within this approach, the leading learning tools of bachelors of electromechanics are mobile Internet devices (MID) – a multimedia mobile devices that provide wireless access to information and communication Internet services for collecting, organizing, storing, processing, transmitting, presenting all kinds of messages and data. The authors reveals the main possibilities of using MID in learning to ensure equal access to education, personalized learning, instant feedback and evaluating learning outcomes, mobile learning, productive use of time spent in classrooms, creating mobile learning communities, support situated learning, development of continuous seamless learning, ensuring the gap between formal and informal learning, minimize educational disruption in conflict and disaster areas, assist learners with disabilities, improve the quality of the communication and the management of institution, and maximize the cost-efficiency. Bachelor of electromechanics competency in modeling of technical objects is a personal and vocational ability, which includes a system of knowledge, skills, experience in learning and research activities on modeling mechatronic systems and a positive value attitude towards it; bachelor of electromechanics should be ready and able to use methods and software/hardware modeling tools for processes analyzes, systems synthesis, evaluating their reliability and effectiveness for solving practical problems in professional field. The competency structure of the bachelor of electromechanics in the modeling of technical objects is reflected in three groups of competencies: general scientific, general professional and specialized professional. The implementation of the technique of using MID in learning bachelors of electromechanics in modeling of technical objects is the appropriate methodic of using, the component of which is partial methods for using MID in the formation of the general scientific component of the bachelor of electromechanics competency in modeling of technical objects, are disclosed by example academic disciplines “Higher mathematics”, “Computers and programming”, “Engineering mechanics”, “Electrical machines”. The leading tools of formation of the general scientific component of bachelor in electromechanics competency in modeling of technical objects are augmented reality mobile tools (to visualize the objects’ structure and modeling results), mobile computer mathematical systems (universal tools used at all stages of modeling learning), cloud based spreadsheets (as modeling tools) and text editors (to make the program description of model), mobile computer-aided design systems (to create and view the physical properties of models of technical objects) and mobile communication tools (to organize a joint activity in modeling).
Dr Donald Bailey starts with introductory material considering the problem of embedded image processing, and how some of the issues may be solved using parallel hardware solutions. Field programmable gate arrays (FPGAs) are introduced as a technology that provides flexible, fine-grained hardware that can readily exploit parallelism within many image processing algorithms. A brief review of FPGA programming languages provides the link between a software mindset normally associated with image processing algorithms, and the hardware mindset required for efficient utilization of a parallel hardware design. The design process for implementing an image processing algorithm on an FPGA is compared with that for a conventional software implementation, with the key differences highlighted. Particular attention is given to the techniques for mapping an algorithm onto an FPGA implementation, considering timing, memory bandwidth and resource constraints, and efficient hardware computational techniques. Extensive coverage is given of a range of low and intermediate level image processing operations, discussing efficient implementations and how these may vary according to the application. The techniques are illustrated with several example applications or case studies from projects or applications he has been involved with. Issues such as interfacing between the FPGA and peripheral devices are covered briefly, as is designing the system in such a way that it can be more readily debugged and tuned. Provides a bridge between algorithms and hardwareDemonstrates how to avoid many of the potential pitfallsOffers practical recommendations and solutionsIllustrates several real-world applications and case studiesAllows those with software backgrounds to understand efficient hardware implementationDesign for Embedded Image Processing on FPGAs is ideal for researchers and engineers in the vision or image processing industry, who are looking at smart sensors, machine vision, and robotic vision, as well as FPGA developers and application engineers.The book can also be used by graduate students studying imaging systems, computer engineering, digital design, circuit design, or computer science. It can also be used as supplementary text for courses in advanced digital design, algorithm and hardware implementation, and digital signal processing and applications.Companion website for the book: www.wiley.com/go/bailey/fpga
To guarantee accurate and continuous monitoring of individual animals at a modern livestock farm, farmers nowadays need reliable and affordable technologies to assist them in performing daily management of tasks. The application of the principles and techniques of process engineering to livestock farming to monitor, model, and manage animal production is called precision livestock farming (PLF). Precision livestock farming seems like the only realistic way to support farmers and other stakeholders in the livestock production chain in the near future while at the same time coping with the rising demand for meat. Precision livestock farming is a series of practices aimed at increasing the farmer’s ability to keep contact with individual animals despite the growing intensification of livestock production. It aims to achieve economically, environmentally, and socially sustainable farming through the observation, behavioral interpretation, and control of the smallest possible group of animals. It enables farmers to reduce operational costs such as expenditures to feed, medication, and energy. Moreover, farmers can use PLF technologies to monitor animal health and welfare to ensure that animals live well and are free of diseases. Precision livestock farming systems aim to translate the output of the technology to useful information to the farmer. Commercial products need a combination of hardware complying with certain technical and safety standards in combination with software, a good user interface, a backup solution to store data, an auto-restart function in case of power failure, manual and help functions, and installers who can install and service the product, etc. Results and potential of PLF technology are mostly unknown to animal scientists, veterinarians, ethologists, etc. due to a lack of collaboration among different disciplines. However, there is no doubt that the combination of new technologies with biology offers great opportunities for the EU in terms of realizing and implementing directives as well as in economic and social terms. A lot of data are already automatically registered by the in-house control computers and collected on a farm computer. In practice, however, the pig farmers hardly use this information. As a result, they miss out on money because deviations in the production process are not noticed or noticed too late. However, the biggest challenge with PLF is to convert this growing amount of data into usable information so that, throughout the day, the farmer can use the relevant information directly to manage operations.
Background and Objectives: Increasing DC loads along with DC nature of distributed energy resources (DERs) raises interest to DC microgrids. Conventional droop/non-droop power-sharing in microgrids suffers from load dependent voltage deviation, slow transient response, and requires the parameters of the loads, system and DERs connection status.Methods: In this paper, a new nonlinear decentralized back-stepping control strategy for voltage control and load sharing of DC islanded microgrids is proposed. The proposed method is robust against the load variations and uncertainty in microgrid parameters and has excellent dynamic and steady-state performance under different operating conditions. The major purpose of the proposed controller is to improve the transient performance of MG with load variations and constant power loads (CPLs). The local controller regulates the terminal voltage of DC-DC converter regarding the local quantities without needs to additional data of other system components.Results: For simplicity, the proposed method is simulated with PSIM software on a DC microgrid with two DGs. Different scenarios are studied to present the performance of the proposed method under different operating conditions.Conclusion: The results indicate the capability of the proposed method for voltage control and load sharing in DC microgrids.
Chlorella vulgaris microalgae biomass was cultivated in brewery secondary effluents and used as a recalcitrant effluent to be valorised energetically by anaerobic digestion process. All previous techniques applied for cellular disruption – autoclave, freeze/heating, ultrasound, microwave - provided either high absorption values and release of reducing sugars in the medium or membrane cells damage, compared to the untreated sample, indicating that the pre-treatment action was effective.
The highest methane production was attained by the autoclave and untreated microalgae assays (samples with less permeabilized cells) while the lowest was provide by the microwaves biomasses pre-treatment: 163-178 mL versus 67 mL CH4. COD removal of 27-29 % and 16 % and TS removal of 28-32 % and 17 % were obtained, respectively. The corresponding methane yield achieved values of 0.04 and 0.030 L g-1 COD and 0.205-0.235 L g-1 TS related to concentrations determined in the influent.
Chemical engineering, Computer engineering. Computer hardware