30793 research outputs found
Sort by
An Intelligent Deep Learning Framework for Identifying and Profiling Darknet Traffic
Full text linkData Availability Statement:
The datasets analyzed during the current study are publicly available from their original repositories cited within the article.The accurate labeling of darknet traffic plays a vital role in real-time cybersecurity systems, as it enables the reliable identification and control of encrypted network applications. State-of-the-art studies have depended mainly on traditional machine learning with public datasets; however, incorporating deep learning (DL) techniques to analyze darknet traffic is still not effectively explored. This paper presented a unique DL-based framework. It integrated discriminative feature selection with an image-based representation of traffic. The work methodology applies the extraction of the most informative features from raw network flows and transforms them into grayscale images, enabling the effective capture of spatial patterns. Those images will be further processed by a hybrid conventional neural network (CNN) and bidirectional long short-term memory (BiLSTM) architecture that leverages the strengths of the CNN in terms of spatial feature extraction, with the modeling of bidirectional temporal dependencies of BiLSTM. For the model testing, two independent encrypted traffic datasets were combined to build a unified and diversified darknet traffic benchmark. The achieved results prove that the proposed hybrid architecture can achieve as high as 89% classification accuracy with an excellent detection and classification capability for darknet traffic. It confirmed a significant performance improvement of the encrypted traffic analysis by integrating feature selection and image-based DL.This research received no external funding
A comprehensive review of machine learning applications in liquid-based cooling solutions of PV/T systems
Full text linkHighlights:
• First systematic review of machine learning in photovoltaic thermal systems.
• Identifies major research gap in hybrid solar energy system modeling approaches.
• Machine learning techniques achieve excellent prediction accuracy in applications.
• Provides systematic guidance for optimal machine learning method selection.
• Nanofluid systems demonstrate superior performance over conventional cooling.Data availability:
Data will be made available on request.This is a PDF of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability. This version will undergo additional copyediting, typesetting and review before it is published in its final form. As such, this version is no longer the Accepted Manuscript, but it is not yet the definitive Version of Record; we are providing this early version to give early visibility of the article. Please note that Elsevier's sharing policy for the Published Journal Article applies to this version, see:
https://www.elsevier.com/about/policies-and-standards/sharing#4-published-journal-article. Please also note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.This paper presents a systematic review of machine learning (ML) applications in liquid-based photovoltaic-thermal (PV/T) systems, a topic that remains largely unaddressed in the existing review literature despite the growing importance of these hybrid systems in renewable energy. A total of 72 publications are analyzed and categorized across three methodological families: artificial neural networks (ANNs), ensemble methods, and other ML techniques. The review is complemented by a patent landscape analysis covering liquid-based PV/T technologies and a critical assessment of the experimental foundations underlying the reviewed ML models.
The analysis reveals that ANNs dominate PV/T modeling at 63% of reviewed studies, with Multilayer Perceptron being the most frequently applied architecture. Ensemble methods, particularly Random Forest and XGBoost, achieve the highest prediction accuracies with R2 values up to 0.999. Across all ML categories, prediction accuracies exceed R2 = 0.95 in most applications, confirming the effectiveness of ML in capturing the complex thermal-electrical interactions characteristic of PV/T systems. Nanofluid-enhanced and phase change material configurations consistently demonstrate significant performance improvements over conventional water cooling.
The review traces the evolution of ML methods in PV/T research from foundational ANN studies in 2012 through recent Transformer-based and reinforcement learning architectures in 2025. Critical research gaps are identified, including prevalent small dataset sizes in most studies, limited experimental validation of nanofluid ML models, and the absence of ML-related patent activity indicating a disconnect between academic research and commercial deployment. Future research directions are proposed covering standardized datasets, transfer learning, IoT integration for real-time control, and explainable AI for engineering interpretation...
Hybrid modelling of heat transfer systems: Combining physics-based and data-driven approaches for improved prediction and extrapolation
Full text linkData availability:
The data that has been used is confidential.Hybrid machine learning-assisted modelling techniques have gained increasing attention recently in many engineering fields. This is due to the challenges associated with pure first-principles and data-driven models, as the former requires deep phenomenological understanding and might become infeasible to describe a complex system with, and the latter needs extensive high-quality data and, more importantly, extrapolates poorly compared to its first principles counterparts. The integration of the two techniques in a framework will result in an integrated approach that benefits from the two realms by strengthening extrapolation capabilities, higher prediction accuracy, and less data demanding and more data-efficient. In this study, a systematic hybrid modelling framework is developed, allowing for the integration of mechanistic models and machine learning algorithms in parallel and series for modelling heat transfer systems to predict a desired target variable, as long as the system is not of a dynamic nature. The framework is developed according to a previous study that enabled the use of machine learning models for such systems. The application of the hybrid modelling framework in this study is demonstrated on the prediction of the condensation heat transfer coefficient in a microfin tube. A laboratory-scale dataset of 5708 datapoints is used for the validation of the developed framework. The validation of the model has been carried out in two different scenarios, both assessing the general prediction and extrapolation capabilities of the developed models in comparison with pure mechanistic and pure machine learning models. The hybrid models, series and parallel, outperform the mechanistic model by approximately 60% more accurate predictions and the machine learning model by 25%, while interpolating. More importantly, while extrapolating, the hybrid models showed approximately 50% more accurate predictions compared to pure machine learning and 27% more accurate compared to the mechanistic model.Hexxcell Ltd
Coalitional Politics and Ethics: Tania Bruguera's Coalitional Work and Artist Collectives
No full text......Brunel University London: BRIEF Award (‘BRUNEL RESEARCH INITIATIVE AND ENTERPRISE FUND’
Knitting and crochet in My Weekly, Woman's Weekly, and Woman's Own, 1914-1918: patriotism, productive leisure, and profit
No full text........
CommGPT: A Graph and Retrieval- Augmented Multimodal Communication Foundation Model
Full text linkThe preprint version of the magazine article is archived on this institutional repository. It is also available at arXiv:2502.18763v1 [cs.IT] (https://arxiv.org/abs/2502.18763 -- [v1] Wed, 26 Feb 2025 02:44:21 UTC (1,089 KB)). It has not been certified by peer review.Large Language Models (LLMs) exhibit advanced cognitive and decision-making capabilities, positioning them as a pivotal technology for 6G networks. However, applying LLMs to the communication domain faces three major challenges: 1) Inadequate communication data; 2) Restricted input modalities; and 3) Difficulty in knowledge retrieval. To overcome these issues, we propose CommGPT, a multimodal foundation model designed specifically for communications. First, we create high-quality pretraining and fine-tuning datasets tailored to communication, enabling the LLM to engage in further pretraining and fine-tuning with communication concepts and knowledge. Then, we design a multimodal encoder to understand and process information from various input modalities. Next, we construct a Graph and Retrieval-Augmented Generation (GRG) framework, efficiently coupling Knowledge Graph (KG) with Retrieval-Augmented Generation (RAG) for multi-scale learning. Finally, we demonstrate the feasibility and effectiveness of the CommGPT through experimental validation
UK Live Comedy Sector Survey Report 2025
No full textThe UK Live Comedy Sector Survey 2025 was jointly conducted by the Centre for Comedy Studies Research at Brunel University, the Live Comedy Association, and British Comedy Guide. The UK Live Comedy Sector Survey was administered by Brunel University of London and ethical approval to conduct the survey was received from the College of Arts, Law and Social Sciences Research Ethics Committee at Brunel University of London.At head of title page: comedysurvey.co.uk .The survey was open from 15th July to 18th August 2025 and from 13th October to 24th October 2025. The survey was distributed through industry press and several further industry distribution lists and networks. Survey data was analysed using descriptive statistics and thematic analysis.This report details the main findings of the UK Live Comedy Sector Survey 2025 run by the Centre for Comedy Studies Research (CCSR), the Live Comedy Association (LCA) and British Comedy Guide (BCG). This is a follow-up to the first Live Comedy Sector Survey, conducted in 2024, and examines the economic, social and cultural impact of the UK live comedy sector. It also provides a progress update on the series of recommendations made in the 2024 report. The 2025 survey was completed by 272 people working in UK live comedy. 63% of respondents were comedians and 37% were people working as either comedy promoters, producers, venue managers, festival organisers, agents, technicians, publishers, journalists or comedy critics. Survey findings and their related recommendations are clustered around 4 key themes: The economics of the live comedy sector; The spaces and places of live comedy; The social impact of live comedy; and inequalities and inequities of live comedy.Live Comedy Association; Brunel University of London. Centre for Comedy Studies Research (CCSR); British Comedy Guide
How does leveraging artificial intelligence in assessments impact student outcomes? a systematic review
Full text linkSupplementary data are available online at: https://www.sciencedirect.com/science/article/pii/S1574013726000389?via%3Dihub#s0255 .Advancements in Artificial Intelligence (AI) are having a profound impact across numerous domains, including education, particularly in the area of assessment. Within higher education, AI-based assessment has gained increasing attention for its potential to enhance student learning processes and outcomes. Following PRISMA guidelines and covering research published between 1997 and 2024, this systematic literature review (SLR) analyzes 159 studies that apply AI techniques, including machine learning (ML), deep learning (DL), and large language models (LLMs), in formative and summative assessment contexts to predict student outcomes. The findings indicate that, while AI integration can enhance assessment strategies and learning outcomes, classification-based models dominate the literature, and more than 80% of studies rely on private or institution-specific datasets, limiting reproducibility and large-scale validation. This review offers a comprehensive comparative synthesis of AI-driven formative and summative assessment approaches in higher education, highlighting methodological trends, evidence, and research gaps
Experimental Study on Ultra-Precision Turning of Freeform Optical Surfaces of Polymethyl Methacrylate with Nanometer Surface Roughness
Full text linkData Availability Statement:
The raw data supporting the conclusions of this article will be made available by the authors on request.Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/app16031350/s1, Figure S1. Flowchart for Assessing the Concavity and Convexity Properties of Free-Form Surfaces.The high performance of optical components is contingent upon the quality of their optical surfaces, thereby imposing elevated standards on the methodologies employed for their fabrication. This study involved experimental research on freeform optical surface elements of polymethyl methacrylate with nano-surface roughness. In this study, the effects of machining parameters of ultra-precision slow tool servo turning on the surface roughness of different types of areas of freeform optical surfaces in the finishing stage were analysed. Based on the analysis of ultra-precision turning test results for freeform optical surfaces, a novel evaluation method for surface quality is proposed to assess the overall uniformity of surface quality across the entire freeform optical surface. Building upon this proposed evaluation method for overall surface quality uniformity, the processing method of high-quality freeform optical surfaces is studied. The results show that in the finishing stage, the radial feed rate exerts the greatest influence on the surface roughness of the freeform optical surface, especially the surface roughness of the concave surface area. This will exacerbate the surface quality inhomogeneity of the freeform optical surface. Based on the analysis results, optimal machining parameters were selected for processing trials. Concurrently, additional machining tests were conducted to further validate the influence of radial feed rate. Ultimately, a nano-scale PMMA freeform optical surface with uniform overall surface quality was achieved. The variation in surface roughness in different regions of the optical freeform is regulated to stabilise within 2 nm on the surface of polymethyl methacrylate. The overall uniformity of surface quality across the entire freeform optical surface was maintained at a high level.This work was supported by National Natural Science Foundation of China (Grant numbers [U2430215])
Quantum-Cognitive Radar: Adaptive Detection with Entanglement under Thermal-Loss Channels
Full text linkCorrespondence.An adaptive Quantum-Cognitive Radar (QCR), which incorporates a two-mode squeezed-vacuum (TMSV) transmitter, a joint idler-signal receiver, and a Quantum Neural Network (QNN) controller to optimize parameters in real time, is introduced through this exchange of correspondence. An expression for a Gaussian correlation detector has been found for thermal-loss channels and compared with the quantum Chernoff bound (QCB). Hardware-aware simulations show that QCR achieves higher detection probability PD at a fixed false-alarm probability PFA (i.e., the probability of declaring a target when it is absent) than both coherent-state radar and nonadaptive quantum baselines. At PFA = 0.05, QCR provides an approximately 3 dB advantage with up to 40% reduction in integration time while maintaining robustness as background noise increases. At the operationally stringent PFA = 10^{−3}, QCR achieves PD = 0.47 versus 0.20 for classical radar, corresponding to a 135% relative improvement. The receiver requires only homodyne/heterodyne sampling and digital correlation, making it compatible with noisy intermediate-scale quantum (NISQ) hardware. The adaptive policy optimizes the parameter vector (M, NS , B, Tint, G) under fixed energy constraints, demonstrating that online adaptation preserves and ex-tends quantum-illumination advantages in nonstationary sensing environments
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇