Neural Network-based Approach Research Articles

A Review of Neural Network Approach on Engineering Drawing Recognition and Future Directions

Engineering Drawing (ED) digitization is a crucial aspect of modern industrial processes, enabling efficient data management and facilitating automation. However, the accurate detection and recognition of ED elements pose significant challenges. This paper presents a comprehensive review of existing research on ED element detection and recognition, focusing on the role of neural networks in improving the analysis process. The study evaluates the performance of the YOLOv7 model in detecting ED elements through rigorous experimentation. The results indicate promising precision and recall rates of up to 87.6% and 74.4%, respectively, with a mean average precision (mAP) of 61.1% at IoU threshold 0.5. Despite these advancements, achieving 100% accuracy remains elusive due to factors such as symbol and text overlapping, limited dataset sizes, and variations in ED formats. Overcoming these challenges is vital to ensuring the reliability and practical applicability of ED digitization solutions. By comparing the YOLOv7 results with previous research, the study underscores the efficacy of neural network-based approaches in handling ED element detection tasks. However, further investigation is necessary to address the challenges above effectively. Future research directions include exploring ensemble methods to improve detection accuracy, fine-tuning model parameters to enhance performance, and incorporating domain adaptation techniques to adapt models to specific ED formats and domains. To enhance the real-world viability of ED digitization solutions, this work highlights the importance of conducting testing on diverse datasets representing different industries and applications. Additionally, fostering collaborations between academia and industry will enable the development of tailored solutions that meet specific industrial needs. Overall, this research contributes to understanding the challenges in ED digitization and paves the way for future advancements in this critical field.

Interpretable Imitation Learning with Symbolic Rewards

Sample inefficiency of deep reinforcement learning methods is a major obstacle for their use in real-world tasks as they naturally feature sparse rewards. In fact, this from-scratch approach is often impractical in environments where extreme negative outcomes are possible. Recent advances in imitation learning have improved sample efficiency by leveraging expert demonstrations. Most work along this line of research employs neural network-based approaches to recover an expert cost function. However, the complexity and lack of transparency make neural networks difficult to trust and deploy in the real world. In contrast, we present a method for extracting interpretable symbolic reward functions from expert data, which offers several advantages. First, the learned reward function can be parsed by a human to understand, verify and predict the behavior of the agent. Second, the reward function can be improved and modified by an expert. Finally, the structure of the reward function can be leveraged to extract explanations that encode richer domain knowledge than standard scalar rewards. To this end, we use an autoregressive recurrent neural network that generates hierarchical symbolic rewards represented by simple symbolic trees. The recurrent neural network is trained via risk-seeking policy gradients. We test our method in MuJoCo environments as well as a chemical plant simulator. We show that the discovered rewards can significantly accelerate the training process and achieve similar or better performance than neural network-based algorithms.

Neural-network-based automatic trajectory adaptation for quality characteristics control in powder compaction

AbstractFuture manufacturing systems will have to become more intelligent to be able to guarantee a constantly high quality of products while simultaneously reducing labor-intensive quality-assurance tasks to address the shortage in workforce. In this work, we study the application of neural networks to the field of powder metallurgy and more specifically the production of green parts as part of a typical sintering process. More specifically, we explore the usage of neural-network-based predictions in closed-loop control. We train neural networks based on a series of produced workpieces, and use these networks in closed-loop production to predict quality characteristics like weight and dimensions of the workpiece in real-time. Based on these predictions an adaptive trajectory planner adjusts then trajectory key points and with this the final piston trajectories to bring and keep quality characteristics of workpieces within tolerance. We finally compare the control performance of this neural network-based approach with a pure sensor-based approach. Results indicate that both approaches are able to bring and keep quality characteristics within their tolerance limits, but that the neural network-based approach outperforms the sensor-based approach in the transient phase, whereas in steady state the neural network needed to be updated from time to time to reach the same high performance as the sensor-based approach. Since updating needs to be performed only from time to time, required expensive sensors can be shared among multiple machines and thus, costs can be reduced. At the same time the superior prediction performance of the neural-network-based approach in transient phases can be exploited to accelerate setting up times for new workpieces. Future work will target the automation of the recording of the training dataset, the exploration of further machine learning methods as well as the integration of additional sensor data to further improve predictions.

A New Deep Neural Network Based on SwinT-FRM-ShipNet for SAR Ship Detection in Complex Near-Shore and Offshore Environments

The advent of deep learning has significantly propelled the utilization of neural networks for Synthetic Aperture Radar (SAR) ship detection in recent years. However, there are two main obstacles in SAR detection. Challenge 1: The multiscale nature of SAR ships. Challenge 2: The influence of intricate near-shore environments and the interference of clutter noise in offshore areas, especially affecting small-ship detection. Existing neural network-based approaches attempt to tackle these challenges, yet they often fall short in effectively addressing small-ship detection across multiple scales and complex backgrounds simultaneously. To overcome these challenges, we propose a novel network called SwinT-FRM-ShipNet. Our method introduces an integrated feature extractor, Swin-T-YOLOv5l, which combines Swin Transformer and YOLOv5l. The extractor is designed to highlight the differences between the complex background and the target by encoding both local and global information. Additionally, a feature pyramid IEFR-FPN, consisting of the Information Enhancement Module (IEM) and the Feature Refinement Module (FRM), is proposed to enrich the flow of spatial contextual information, fuse multiresolution features, and refine representations of small and multiscale ships. Furthermore, we introduce recursive gated convolutional prediction heads (GCPH) to explore the potential of high-order spatial interactions and add a larger-sized prediction head to focus on small ships. Experimental results demonstrate the superior performance of our method compared to mainstream approaches on the SSDD and SAR-Ship-Dataset. Our method achieves an F1 score, mAP0.5, and mAP0.5:0.95 of 96.5% (+0.9), 98.2% (+1.0%), and 75.4% (+3.3%), respectively, surpassing the most competitive algorithms.

Dynamic pricing of differentiated products under competition with reference price effects using a neural network-based approach

Dynamic pricing of differentiated products under competition with reference price effects using a neural network-based approach

Neural Network-Based Approach to Detect and Filter Misleading Audio Segments in Classroom Automatic Transcription

Audio recording in classrooms is a common practice in educational research, with applications ranging from detecting classroom activities to analyzing student behavior. Previous research has employed neural networks for classroom activity detection and speaker role identification. However, these recordings are often affected by background noise that can hinder further analysis, and the literature has only sought to identify noise with general filters and not specifically designed for classrooms. Although the use of high-end microphones and environmental monitoring can mitigate this problem, these solutions can be costly and potentially disruptive to the natural classroom environment. In this context, we propose the development of a novel neural network model that specifically detects and filters out problematic audio sections in classroom recordings. This model is particularly effective in reducing transcription errors, achieving up to a 96% success rate in filtering out segments that could lead to incorrect automated transcriptions. The novelty of our work lies in its targeted approach for low-budget, aurally complex environments like classrooms, where multiple speakers are present. By allowing the use of lower-quality recordings without compromising analysis capability, our model facilitates data collection in natural educational settings and reduces the dependency on expensive recording equipment. This advancement not only demonstrates the practical application of specialized neural network filters in challenging acoustic environments but also opens new avenues for enhancing audio analysis in educational research and beyond.

Neural network-based approach for identifying fake news

Neural network-based approach for identifying fake news

Neural network-based approach for identifying fake news

Neural network-based approach for identifying fake news

Retracted: A Deep Neural Network-Based Approach to Media Hotspot Discovery

Retracted: A Deep Neural Network-Based Approach to Media Hotspot Discovery

Feature Reduction Networks: A Convolution Neural Network-Based Approach to Enhance Image Dehazing

Image dehazing represents a dynamic area of research in computer vision. With the exponential development of deep learning, particularly convolutional neural networks (CNNs), innovative and effective image dehazing techniques have surfaced. However, in stark contrast to the majority of computer vision tasks employing CNNs, the output from a dehazing model is often treated as uninformative noise, even though the model’s filters are engineered to extract pertinent features from the images. The standard approach of end-to-end models for dehazing involves noise removal from the hazy image to obtain a clear one. Consequently, the model’s dehazing capacity diminishes, as the noise is progressively filtered out throughout the propagation phase. This leads to the conception of the feature reduction network (FRNet), which is a distinctive CNN architecture that incrementally eliminates informative features, thereby resulting in the output of noise. Our experimental results indicate that the CNN-driven FRNet surpasses previous state-of-the-art (SOTA) methods in terms of the peak signal-to-noise ratio (PSNR) and structural similarity index measure (SSIM) evaluation metrics. This highlights the effectiveness of the FRNet across various image dehazing datasets. With its reduced overhead, the CNN-based FRNet demonstrates superior performance over current SOTA methods, thereby affirming the efficacy of CNNs in image dehazing tasks.

Optimizing Automatic Voltage Regulation: A Deep Reinforcement Learning Approach

This paper presents a novel approach to enhance the performance of Automatic Voltage Regulator (AVR) systems in power systems using Deep Reinforcement Learning (DRL). The AVR plays a critical role in maintaining voltage stability and ensuring reliable power delivery. However, conventional control strategies, such as PID controllers, have limitations in handling complex and nonlinear power system dynamics. In this study, the application of DRL techniques is explored, particularly the Twin-Delayed Deep Deterministic Policy Gradient (TD3) algorithm, to AVR control. This algorithm offers the advantage of handling continuous action spaces and enable the controller to learn optimal control policies directly from the system's state information. The results show that the DRL approach outperforms the traditional PID and Neural Network-based control approaches, with the shortest time response and the best voltage regulation performance. The use of DRL in AVR system control shows promising potential for improving the efficiency and accuracy of power system control. This research provides insights into the advantages of DRL for process control and highlights its potential for future applications in power system control.

1D Convolutional Neural Network-based Signal Tracking Algorithm with Feature Extraction and System Diagnosis

Diverse adaptive algorithms have been investigated and developed in an effort to mitigate the vibration and noise produced by mechanical systems. A reference signal plays a crucial role in adaptive algorithms for monitoring the target signal, however it is difficult to identify the reference signal when its spectrum is relatively complex. In addition, if the state of a mechanical system changes due to a disturbance or malfunction, signal tracking may not perform as well if the reference signal for the changing condition has not been redefined. This work focuses on constructing a signal tracking mechanism based on deep learning to increase tracking performance. The suggested algorithm consists of the following two parts: 1) automatically defining the reference signal segment and 2) system state feedback segment. These portions are managed by convolutional neural networks in one dimension. The Case Western Reserve University (CWRU) bearing dataset was used for training and testing to validate the proposed technique. In addition, a neural network-based approach for signal tracking is applied to validate the signal tracking performance. The suggested approach demonstrates that the signal tracking performance is great regardless of the system's state.

A Network Representation Learning Model Based on Multiple Remodeling of Node Attributes

Current network representation learning models mainly use matrix factorization-based and neural network-based approaches, and most models still focus only on local neighbor features of nodes. Knowledge representation learning aims to learn low-dimensional dense representations of entities and relations from structured knowledge graphs, and most models use the triplets to capture semantic, logical, and topological features between entities and relations. In order to extend the generalization capability of the network representation learning models, this paper proposes a network representation learning algorithm based on multiple remodeling of node attributes named MRNR. The model constructs the knowledge triplets through the textual association relationships between nodes. Meanwhile, a novel co-occurrence word training method has been proposed. Multiple remodeling of node attributes can significantly improve the effectiveness of network representation learning. At the same time, MRNR introduces the attention mechanism to achieve the weight information for key co-occurrence words and triplets, which further models the semantic and topological features between entities and relations, and it makes the network embedding more accurate and has better generalization ability.

An artificial neural network-based machine learning approach to correct coarse-mesh-induced error in computational fluid dynamics modeling of cell culture bioreactor

Computational Fluid Dynamics (CFD) is a valuable tool for studying fluid environments within cell culture bioreactors and optimizing processing parameters, but it can be computationally expensive. This study developed an artificial neural network (ANN)-based machine learning model to predict and correct the coarse-mesh-induced errors in CFD modeling of a spinner flask bioreactor. A baseline ANN model was trained to predict the velocity error function between the coarse and optimized reference mesh results at one rotational speed (90 rpm), demonstrating that the ANN-based approach could correct the coarse-mesh velocity with RMSE values of nodal velocities improved by an average of ∼20% at different rotational speeds. The effect of ANN structure, input data normalization, and training dataset combinations on prediction performance was evaluated. More neurons and hidden layers generated better results but required more computational time for training. The model’s generalization capabilities were further evaluated in case studies of correcting velocity and Kolmogorov length at different fluid viscosity and bioreactor impeller geometry conditions. Results suggested that the ANN model had better generalization in correcting Kolmogorov length than velocity. This research provides insights into using a machine learning approach to enhance CFD modeling in bioreactor applications, contributing to advancing tissue engineering processes.

Real-time gypsum quality estimation in an industrial calciner: A neural network-based approach

Total bound moisture (TBM) is a typical quality indicator of industrial-grade gypsum. This gypsum is comprised of three distinct phases, namely anhydrite, dihydrate, and hemihydrate, of which only the latter is of much industrial use. TBM analysis is a lengthy laboratory procedure, and an artificial neural network (ANN) TBM inference measurement is proposed as a fast and online alternative. An ANN inference model for gypsum TBM based on plant data was developed. The inputs to the network were primarily focused on the plant's calciner, and different network topologies, data divisions, and transfer functions were investigated. Furthermore, the applicability of the TBM value as a quality indicator was investigated based on a gypsum phase analysis. A strong correlation between TBM and the gypsum hemihydrate and anhydrite content was found, validating the plant target TBM of 5.8% as a quality indicator. A network topology consisting of one hidden layer with logarithmic-sigmoid (logsig) and pure linear (purelin) transfer functions showed the best performance (R > 90%).

A novel transfer learning-based model for diagnosing malaria from parasitized and uninfected red blood cell images

Malaria represents a potentially fatal communicable illness triggered by the Plasmodium parasite. This disease is transmitted to humans through the bites of Anopheles mosquitoes that carry the infection. This disease has significant and devastating consequences on the health systems of fragile countries, particularly in sub-Saharan Africa. Malaria affects red blood cells by invading and replicating within them, destroying them, and releasing toxic byproducts into the bloodstream. The parasite’s ability to stick and modify the surface of red blood cells can cause them to become sticky, obstructing blood flow in vital organs such as the brain and spleen. Therefore, efficient approaches for the early detection of malaria are critical to saving patients’ lives. The main aim of this study is to develop an efficient model for early malaria diagnosis. We used malaria images based on parasitized and uninfected red blood cells for the study experiments. We applied neural network-based approaches such as Neural Search Architecture Network (NASNet) and compared its performance with machine learning techniques. Moreover, we proposed a novel NNR (NASNet Random forest) method for feature engineering. The proposed NNR approach first extracts spatial features from input malaria images, then class prediction probability features are extracted from these spatial features. The feature set obtained from the data extraction trains machine learning models. Our comprehensive experiments show that the support vector machine outperformed state-of-the-art models, achieving a high-performance score of 99% and having an inference time near 0.025 s. We validated the performance using k-fold cross-validation and optimized the hyperparameters through tuning. Our proposed research has improved the early diagnosis of malaria and can assist medical specialists in reducing the mortality rate.

Diffusion Convolution Neural Network-Based Multiview Gesture Recognition for Athletes in Dynamic Scenes

This paper focuses on deep vision sensing-assisted gesture recognition for athletes in dynamic scenes. Although many research attention had been devoted to this field in recent years, most of existing works failed to fully take characteristics of dynamic scenes into consideration. To deal with this challenge, this paper proposes a diffusion convolution neural network-based multiview gesture recognition approach in dynamic scenes. For one thing, the dynamic spatiotemporal slice position selection based on the body mask heatmap is adopted to calculate positions of horizontal and vertical slices. Thus, the dynamic selection of slice positions in two directions can be realized, and then the extraction of bi-directional spatiotemporal slice images can be completed. For another, action sequences through the 3D residual neural network are learned, and the spatiotemporal information among frames are mined through recurrent networks. Through their combination, a multi-view gesture recognition approach for athletes is constructed. In the experiments, two standard datasets UCF101 and HMDB51 are utilized to establish simulation environment. The proposed method can reach the accuracy beyond 95% on the two datasets. Compared with several typical recognition methods, the proposed method shows higher accuracy.

Health Monitoring and Diagnosis for Geo-Distributed Edge Ecosystem in Smart City

With the increasing number of Internet of things (IoT) devices being deployed and used in daily life, the load on computational devices has grown exponentially. This situation is more prevalent in smart cities where such devices are used for autonomous control and monitoring. Smart cities have different kinds of applications that are aided through IoT devices that collect data, send it to computational processing and storage devices, and get back decisions or actuate the actions based on the input data. There has been a stringent requirement to reduce the end-to-end delay in this process owing to the remote deployment of cloud data centres. This eventually led to the revolution of edge computing, wherein nano-micro-processing devices can be deployed closer to the premises of the smart application and process the data generated with a lower turnaround time. However, due to the limited computational power and storage, controlling the workload diverted to the edge devices has been challenging. The workload scheduling policies and task allocation schemes often fail to consider the run time health of the edge devices due to a lack of proper monitoring infrastructure. Thus, in this paper, we proposed a health monitoring and diagnosis framework for geo-distributed edge clusters processing big data generated by smart city applications. This framework is built over the Map-Reduce approach for distributed processing of big data on edge clusters deployed across the smart city. Within this framework, SmartMonit (a monitoring agent) is deployed that collects the health statistics of edge devices and predicts the potential failures using an artificial neural network-based self-organising maps approach. The proposed framework is deployed over different clusters to test the efficacy concerning failure detection.

Scalable Polar Code Construction for Successive Cancellation List Decoding: A Graph Neural Network-Based Approach

Scalable Polar Code Construction for Successive Cancellation List Decoding: A Graph Neural Network-Based Approach

Localizing False Data Injection Attacks in Smart Grid: A Spectrum-Based Neural Network Approach

Smart grid is confronted with cyberattacks due to the increasing dependence on cyberspace. False data injection attacks (FDIAs) represent a major type of cyberattacks that cannot be detected by the traditional bad data detection (BDD). The majority of existing researches focus on how to detect FDIAs, while little attention has been paid to obtaining the exact attack locations, which are crucial for deploying countermeasures in time. In this paper, we propose a neural network-based approach for online localizing FDIAs in AC power systems. The approach transforms the multi-dimensional time series measurements to their time-frequency and spectral graph representations, to jointly capture the temporal and spatial correlations in the spectral domain. The spectral decomposition could better characterize the latent properties of measurements, such as auto-correlations of a bus and key dependency between buses, which would be poorly represented in the time domain and vertex domain. Specifically, the proposed approach first adopts the short-time Fourier transform (STFT) and a two-channel convolutional neural network (2C-CNN) to model the spectral-temporal correlations. Then the spectral-spatial relationships are modeled with a graph convolutional network (GCN), which combines the nodal admittance matrix and physical property of power systems. The prior physical knowledge further improves our approach’s interpretability and localization performance. We evaluate the effectiveness of our proposed approach with comprehensive case studies on various standard test systems. Numerical results indicate the superiority of our approach over baselines.