Recursive Learning Research Articles

A Community Perspective on Boxing, Well‐being and Young People

ABSTRACTCommunities possess various resources, such as sporting organisations, that establish positive relationships with young people whom statutory services may struggle to engage. Boxing has been shown to attract individuals from diverse cultural and socio‐economic backgrounds, offering proven benefits in areas such as physical health and engagement. When taking into consideration mental and physical well‐being, there is limited knowledge regarding the overall well‐being benefits of boxing for young people. This study adopts an ethnographic approach within an Inner London community boxing gym to investigate the well‐being benefits of boxing for young people. Qualitative interviews, field notes and observations were employed with young people, staff and volunteers. The community identified that boxing provided young people with a changed sense of self, engaged them in a recursive learning process and instilled a sense of belonging, like that of a new family. These factors can all be associated with well‐being. The study highlights the significance of community‐based resources, such as boxing gyms, in providing support for improving well‐being for a diverse cohort of young people. Please refer to the Supplementary Material section to find this article's Community and Social Impact Statement

D2-SPDM: Faster R-CNN-Based Defect Detection and Surface Pixel Defect Mapping with Label Enhancement in Steel Manufacturing Processes

The steel manufacturing process is inherently continuous, meaning that if defects are not effectively detected in the initial stages, they may propagate through subsequent stages, resulting in high costs for corrections in the final product. Therefore, detecting surface defects and obtaining segmentation information is critical in the steel manufacturing industry to ensure product quality and enhance production efficiency. Specifically, segmentation information is essential for accurately understanding the shape and extent of defects, providing the necessary details for subsequent processes to address these defects. However, the time-consuming and costly process of generating segmentation annotations poses a significant barrier to practical industrial applications. This paper proposes a cost-efficient segmentation labeling framework that combines deep learning-based anomaly detection and label enhancement to address these challenges in the steel manufacturing process. Using ResNet-50, defects are classified, and faster region convolutional neural networks (faster R-CNNs) are employed to identify defect types and generate bounding boxes indicating the defect locations. Subsequently, recursive learning is performed using the GrabCut algorithm and the DeepLabv3+ model based on the generated bounding boxes, significantly reducing annotation costs by generating segmentation labels. The proposed framework effectively detects defects and accurately defines them, even in randomly collected images from the steel manufacturing process, contributing to both quality control and cost reduction. This study presents a novel approach for improving the quality of the steel manufacturing process and is expected to enhance overall efficiency in the steel manufacturing industry.

Three-Stage Recursive Learning Technique for Face Mask Detection on Imbalanced Datasets

Three-Stage Recursive Learning Technique for Face Mask Detection on Imbalanced Datasets

CryptoGenSec: A Hybrid Generative AI Algorithm for Dynamic Cryptographic Cyber Defence

As the world of cybersecurity constantly changes, traditional cryptographic techniques have faced limitations in the context of today's sophisticated and dynamic threats. Existing protections usually adopt static algorithms and key structures, making it difficult for them to resist the categories of modern attacks. This research paper, therefore, presents CryptoGenSec, a brand-new generative AI algorithm based on a hybrid consisting of generative adversarial networks (GANs) on reconnaissance learning (RL), for the purpose of increasing cryptographic cyber defences. CryptoGenSec applies a GAN to simulate various types of attack scenarios in cyberspace to perceive possible vulnerabilities. Then, RL refines the response strategies of our algorithm through recursive learning from the above simulations in real time and realizes the dynamic adaptation and evolution of defense mechanisms. By assessing the results of CryptoGenSec’s performance when traditional security methods are used as baselines, we can use several metrics for evaluation, such as detection accuracy, response time, resilience and evolution ability. According to these findings, the superiority of CryptoGenSec over conventional mechanisms becomes evident. To be more specific, it even shows an overwhelming edge in terms of threat detection, resulting in a 20% increase in speed of response, a 30% decrease in speed of response, and resisting power, making it 25% harder than the other methods. Moreover, it has a greater possibility of eliminating false-positive effects, which usually come from new and even dawned jeopardy: 50%. Moreover, to highlight the making-a-fortune frauds in the zero-day world, a comparison of the cohorts makes CryptoGenSec a 40% upper step. Stopping attackers from taking away all their data is also its plus point, which gains 95% achievement, whereas using mere methods only results in a 70% possibility. An enormous step in cybersecurity was taken with the combination of GANs and RL within the CryptoGenSec algorithm. Instead of being defenceless against all attacks, this approach changes and matches the threat level when necessary. The highly promising results presented here demonstrate its potential as a crucial technology for addressing the growing complexities of cyber challenges. This is a large step toward making defensive mechanisms more efficient and reliable.

Recursive Learning Based Smart Energy Management With Two-Level Dynamic Pricing Demand Response

Recursive Learning Based Smart Energy Management With Two-Level Dynamic Pricing Demand Response

Machine collaboration

We propose a new ensemble framework for supervised learning, called machine collaboration (MaC), using a collection of possibly heterogeneous base learning methods (hereafter, base machines) for prediction tasks. Unlike bagging/stacking (a parallel and independent framework) and boosting (a sequential and top‐down framework), MaC is a type of circular and recursive learning framework. The circular and recursive nature helps the base machines to transfer information circularly and update their structures and parameters accordingly. The theoretical result on the risk bound of the estimator from MaC reveals that the circular and recursive feature can help MaC reduce risk via a parsimonious ensemble. We conduct extensive experiments on MaC using both simulated data and 119 benchmark real datasets. The results demonstrate that in most cases, MaC performs significantly better than several other state‐of‐the‐art methods, including classification and regression trees, neural networks, stacking, and boosting.

RRTrN: A lightweight and effective backbone for scene text recognition

Models based on extremely deep convolutional networks and attention mechanisms now have powerful feature extraction abilities, improving text recognition performance in natural scenes. However, the very deep network model built with stacked layers brings massive parameters, which limits the application of the text recognition algorithm on storage-constrained devices. In this paper, we propose a lightweight and effective backbone called the Recursive Residual Transformer Network (RRTrN) for scene text recognition. Specifically, by leveraging recursive learning and a combination of convolutional layers and a transformer unit, RRTrN achieves powerful feature extraction while significantly reducing the number of parameters. This reduction in parameters promotes the deployment of our text recognition algorithm on storage-constrained devices, making it more accessible for practical applications. Furthermore, a recursive distillation strategy is presented to balance the recursive learning inference time and performance, enhancing the practicality and efficiency of RRTrN. Extensive experiments on mainstream benchmarks and popular models verify the generalization of RRTrN and achieve state-of-the-art recognition performance on five datasets. Notably, the classical STR model based on RRTrN can achieve a 3 percentage point increase in recognition accuracy or reduce the number of parameters by 80%.

Toward an organism-oriented ontology in organizing health, or recursive learning as resistance.

Toward an organism-oriented ontology in organizing health, or recursive learning as resistance.

Finite-Horizon $H_{\infty }$ Consensus Control of Discrete Time-Varying Multiagent Systems: A Dynamic Event-Based Learning Approach

This paper addresses the finite-horizon <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$H_{\infty }$</tex-math></inline-formula> consensus problem for general discrete time-varying multi-agent systems with energy-bounded external disturbances and limited network bandwidth resource. A dynamic event-based learning approach is proposed to produce a dynamic discrete-time event-triggered scheme and a dynamic distributed <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$H_{\infty }$</tex-math></inline-formula> consensus protocol. Based on the local state information, the dynamic distributed <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$H_{\infty }$</tex-math></inline-formula> consensus protocol is developed to guarantee the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$H_{\infty }$</tex-math></inline-formula> consensus of the general discrete time-varying multi-agent systems with external disturbances. Besides, a novel criterion is established by employing the Lyapunov theory and considering the special dynamic triggering constraint, and the corresponding recursive learning algorithm is developed. In contrast with the previous works, the proposed dynamic event-based approach results in less network bandwidth utilization, and the novel recursive algorithm brings less conservatism of the designed controller. Finally, the effectiveness of the corresponding algorithm and the advantages of the proposed dynamic event-based approach are testified by a numerical simulation.

DIGITAL TECHNOLOGY BASED BUSINESS MODEL DESIGN ANALYSIS

This research investigates how digital technology aid in the development of new business models in the creative industries. A holistic business model framework was built through a thorough literature study, which was then utilized to assess empirical evidence from the creative industries. According to the findings, digital technology have permitted widespread changes in company structures, and several noteworthy patterns have emerged. However, the redesigned business models are frequently not 'new' in the sense of being unprecedented. Business model innovation primarily manifests itself in the use of digital technology to enable the deployment of a broader choice of business models than was previously available to a corporation. The increasing adoption of numerous business models as a portfolio inside one corporation is a key rising trend. This occurs in enterprises of all sizes when a single firm employs numerous business models to service distinct market groups, offer different products, or engage in multi-sided markets, or employs changing business models over time. A recursive learning process refines and extends the holistic business model framework, which can serve as a cognitive instrument for comprehending business models as well as a planning tool for business model improvements. The study advances our understanding of business model theory and how digital technologies enable business model innovation in the creative industries. Three new study topics for the future are emphasized.

Recursive Identification of Time-Varying Hammerstein Systems With Matrix Forgetting

The real-time estimation of the time-varying Hammerstein system by using a noniterative learning schema is considered and extended to incorporate a matrix forgetting factor. The estimation is cast in a Variational-Bayes framework to best emulate the original posterior distribution of the parameters within the set of distributions with feasible moments. The recursive concept we propose approximates the exact posterior comprising undistorted information about the estimated parameters. In many practical settings, the incomplete model of parameter variations is compensated by forgetting of obsolete information. As a rule, the forgetting operation is initiated by the inclusion of an appropriate prediction alternative into the time update. It is shown that the careful formulation of the prediction alternative, which relies on Bayesian conditioning, results in partial forgetting. The paper inspects two options with respect to the order of the conditioning in the posterior, which proves vital in the successful localization of the source of inconsistency in the data-generating process. The geometric mean of the discussed alternatives then modifies recursive learning through the matrix forgetting factor. We adopt the decision-making approach to revisit the posterior uncertainty by dynamically allocating the probability to each of the prediction alternatives to be combined.

Transfer-recursive-ensemble learning for multi-day COVID-19 prediction in India using recurrent neural networks

The COVID-19 pandemic has put a huge challenge on the Indian health infrastructure. With a larger number of people getting affected during the second wave, hospitals were overburdened, running out of supplies and oxygen. Hence, predicting new COVID-19 cases, new deaths, and total active cases multiple days in advance can aid better utilization of scarce medical resources and prudent pandemic-related decision-making. The proposed method uses gated recurrent unit networks as the main predicting model. A study is conducted by building four models pre-trained on COVID-19 data from four different countries (United States of America, Brazil, Spain, and Bangladesh) and fine-tuned on India’s data. Since the four countries chosen have experienced different types of infection curves, the pre-training provides a transfer learning to the models incorporating diverse situations into account. Each of the four models then gives 7-day ahead predictions using the recursive learning method for the Indian test data. The final prediction comes from an ensemble of the predictions of the different models. This method with two countries, Spain and Bangladesh, is seen to achieve the best performance amongst all the combinations as well as compared to other traditional regression models.

Memory Recurrent Elman Neural Network-Based Identification of Time-Delayed Nonlinear Dynamical System

In this article, an attempt has been made to propose an improved version of the classical Elman neural network (ENN) and its application is presented to identify the unknown dynamics of time-delayed nonlinear plants. The model proposed, known as memory recurrent ENN (MRENN), consists of an additional number of weighted self-feedback loops, an extra output context layer, and weighted connections of input signals (through adjustable weights) to the output-layer neuron. The model is proposed to be given only two signals as its inputs: plant unit-delayed value and the present value of the externally applied signal irrespective of the actual order of the plant (which most of the time may not be known). To guarantee stability, the parameters of the MRENN model are updated using the equations that are obtained by applying Lyapunov-stability criteria. Furthermore, the recursive learning rate scheme is constructed for speeding up the learning process. From the simulation results, MRENN model appears to have produced comparably superior outcomes when the performance of the suggested model is compared with that of other well-known models.

Traffic flow monitoring in software-defined network using modified recursive learning

The recursive network design used in this paper to monitor traffic flow ensures accurate anomaly identification. The suggested method enhances the effectiveness of cyber attacks in SDN. The suggested model achieves a remarkable attack detection performance in the case of distributed denial-of-service (DDoS) attacks by preventing network forwarding performance degradation. The suggested methodology is designed to teach users how to match traffic flows in ways that increase granularity while proactively protecting the SDN data plane from overload. The application of a learnt traffic flow matching control policy makes it possible to obtain the best traffic data for detecting abnormalities obtained during runtime, improving the performance of cyber-attack detection. The accuracy of the suggested model is superior to the MMOS, FMS, DATA, Q-DATA, and DEEP-MC by 19.23%, 16.25%, 47.61%, 16.25%, and 12.04%.

Lightweight Semi-supervised Network for Single Image Rain Removal

Deep learning technologies have shown their advantages in Single Image Rain Removal (SIRR) tasks. However, the derained results of most methods are limited to some challenges. First, due to the lack of real-world rainy/clean image pairs, many methods seriously rely on the labeled synthetic training images and will not effectively remove complex rain streaks in real-world scenarios. Second, most existing SIRR models require high computing power, which considerably limits their real-world applications. To address these issues, we propose a Lightweight Semi-supervised Network (LSNet) for SIRR. Our LSNet utilizes a compact semi-supervised framework to improve generalization ability in real-world rainy images removal. Meanwhile, in our semi-supervised framework, we also design a cascaded sub-network, which progressively removes complex rain streaks via a multi-stage manner. Specially, the multi-stage manner is based on a series of cascaded blocks, where we conduct recursive learning strategy to reduce model parameters. Extensive experimental results demonstrate that our method achieves comparable performance to the state-of-the-arts while has fewer parameters.

Online feedforward parameter learning with robustness to set-point variations

High-tech motion system development is driven by increasingly accurate and fast positioning requirements. Feedforward compensation together with high bandwidth feedback control are essential to achieve these ever tightening performance demands. In particular, online adaptation of the feedforward parameters, to correct for small position dependencies and slow variations, is crucial to approach zero error tracking. The aim of this paper is a framework that provides robust recursive learning of feedforward parameters for any bounded reference trajectory. The convergence of the parameter learning strategy exploits the difference in time-scale between the parameter variation rate and the bandwidth of the servo controlled system. This enables to describe a servo-error-based objective function for varying trajectories as a static sector bounded nonlinearity. Subsequently, the circle criterion is employed to derive stability guarantees on the learning with explicit robustness to reference trajectory variation. A numerical case study demonstrates that a significant performance improvement can be achieved.

H-HOME: A learning framework of federated FANETs to provide edge computing to future delay-constrained IoT systems

In 6G systems, it will be mandatory that the network is able to support edge computing powered by Artificial Intelligence (AI) to provide mobile devices with the opportunity of job offloading for computation, so implementing the new paradigm of Intelligent Internet of Intelligent Things (IIoIT). In areas that are very difficult to be covered by the structured networks, Flying Ad-Hoc Networks (FANET) can be considered one of the most promising technologies to enhance coverage, capacity, reliability, and energy efficiency of wireless cellular networks, also providing edge-computing services. The goal of this paper is to propose a two-layer Hierarchical Horizontal-Offload ManagEment (H-HOME) framework for horizontal offload among the Unmanned Aerial Vehicles (UAV) of the same FANET, in order to minimize processing delay and jitter. The framework exploits Federated Reinforcement Learning in order to take advantage of knowledge sharing without incurring into problems of privacy and network overloading. A Markov Decision Process (MDP) is also defined to optimize decisions of the FANET Orchestrator (FO). Simulation results, obtained from two different analyses, demonstrate that H-HOME outperforms the traditional local training approach, based on simple recursive learning, and it can be effectively used to reduce power consumption and increase FANETs flight autonomy.

Two-Dimensional Spatial Construction for Online Modeling of Distributed Parameter Systems

A 2-D spatial construction method is proposed for the online modeling of distributed parameter systems (DPSs), such as battery thermal process. The proposed method can combine the advantages of spectral method and Karhunen–Loève decomposition (KLD) method. First, the continuous spatial basis functions are designed by the 2-D spatial construction to keep the information between sensing locations. With the 2-D space-time separation and recursive learning, the derived model can preserve the couplings between spatial dimensions and update over time. The radial basis function network is utilized to identify the low-dimensional temporal dynamics. After the space-time synthesis, the constructed spatiotemporal model can provide continuous modeling of the DPS with satisfactory performance. Convergence analysis has been carried out, which proves that the proposed method can guarantee bounded errors. Finally, simulations and experiments on a pouch-type lithium-ion battery with unknown partial differential equations prove the effectiveness of the proposed method.

Image Reconstruction Using Deep Neural Networks Models

Abstract: Image restoration is the process of restoring the original image. It can be challenging to eliminate image blur in a variety of contexts, including photography, radar imaging, and the removal of motion blur brought on by camera shaking. Image noise is unintentional signal that enters an image from a sensor,such as a thermal or electrical signal or an external factor like rain or snow. The image degradation may be caused by transmission noise, object motion, resolution restrictions, coding artefacts, camera shake, or a combination of these factors. In order to distinguish between HF and LF artefacts, image decomposition is employed to divide the deformed image into a texture layer and a structure layer (Low Frequency LF Component) The current approach utilises the frequency characteristics of various forms of artefactsthrough a configurable deep neural network structure. Therefore, by changing the architecture, the same method may be applied to a number of picture restoration tasks. A quality enhancement network that uses residual and recursive learning is suggested for decreasing the artefacts with comparable frequency characteristics. Residuallearning is used to enhance performance and speed up the training process. Recursive learning is used to both improve performance and drastically cut down on the amount of training parameters. This Project aims to build systems for reconstructing the old images from under sampled one and mismatched Pixels to form a proper image to increase its visible quality and its pixels quality by using a Deep Neural network Models and it can improve the integration of various feature representations from many photos. Result Shows Improved Training accuracy of 92%.When compared to the two-frame designs now in use, the multi-frame architecture will be used which prevents repetitive computations caused by multiple inferences when aligning multiple images

A new ensemble spatio-temporal PM2.5 prediction method based on graph attention recursive networks and reinforcement learning

Inhalable particulate matter with a diameter of less than 2.5 μm spatio-temporal prediction technology is an important tool for environmental governance in urban traffic congestion areas. A new Ensemble Graph Attention Reinforcement Learning Recursive Network is proposed to create a multi-data-driven spatio-temporal prediction method with excellent application value. The modeling process includes three basic steps. In step I, the graph attention network is used to effectively aggregate the spatio-temporal correlation characteristics of the original air pollutant data. In step II, the features extracted from the graph attention network are transferred to the long short-term memory network and the temporal convolutional network and the prediction models are constructed respectively. In step III, the reinforcement learning algorithm effectively analyzes the adaptability of the two different models to the data sets and realizes ensemble based on continuous optimization of weights. By comparing the experimental results of the listed cases, the following points can be summarized: (a) the graph attention network can effectively aggregate the spatio-temporal correlation characteristics of the original data and optimize the performance of the predictor. (b) The reinforcement learning algorithm effectively realizes the integration of several neural networks and improves the comprehensive adaptability and generalization capabilities of the model. (c) The proposed model in this paper has great application potential and value in spatial and temporal prediction and has achieved better performance than the other 25 benchmark models.