Design Of Neural Network Architectures Research Articles

A genetic programming approach to the automated design of CNN models for image classification and video shorts creation

Neural architecture search (NAS) is a rapidly growing field which focuses on the automated design of neural network architectures. Genetic algorithms (GAs) have been predominantly used for evolving neural network architectures. Genetic programming (GP), a variation of GAs that work in the program space rather than a solution space, has not been as well researched for NAS. This paper aims to contribute to the research into GP for NAS. Previous research in this field can be divided into two categories. In the first each program represents neural networks directly or components and parameters of neural networks. In the second category each program is a set of instructions, which when executed, produces a neural network. This study focuses on this second category which has not been well researched. Previous work has used grammatical evolution for generating these programs. This study examines canonical GP for neural network design (GPNND) for this purpose. It also evaluates a variation of GP, iterative structure-based GP (ISBGP) for evolving these programs. The study compares the performance of GAs, GPNND and ISBGP for image classification and video shorts creation. Both GPNND and ISBGP were found to outperform GAs, with ISBGP producing better results than GPNND for both applications. Both GPNND and ISBGP produced better results than previous studies employing grammatical evolution on the CIFAR-10 dataset.

A Novel Dynamic Operation Optimization Method Based on Multiobjective Deep Reinforcement Learning for Steelmaking Process.

This article studies a dynamic operation optimization problem for a steelmaking process. The problem is defined to determine optimal operation parameters that bring smelting process indices close to their desired values. The operation optimization technologies have been applied successfully for endpoint steelmaking, but it is still challenging for the dynamic smelting process because of the high temperature and complex physical and chemical reactions. A framework of deep deterministic policy gradient is applied to solve the dynamic operation optimization problem in the steelmaking process. Then, an energy-informed restricted Boltzmann machine method with physical interpretability is developed to construct the actor and critic networks in reinforcement learning (RL) for dynamic decision-making operations. It can provide a posterior probability for each action to guide training in each state. Furthermore, in terms of the design of neural network (NN) architecture, a multiobjective evolutionary algorithm is used to optimize the model hyperparameters, and a knee solution strategy is designed to balance the model accuracy and complexity of neural networks. Experiments are conducted on real data from a steelmaking production process to verify the practicability of the developed model. The experimental results show the advantages and effectiveness of the proposed method compared with other methods. It can meet the requirements of the specified quality of molten steel.

Automating the Human Action of First-Trimester Biometry Measurement from Real-World Freehand Ultrasound

Objective: Automated medical image analysis solutions should closely mimic complete human actions to be useful in clinical practice. However, more often an automated image analysis solution only represents part of a human task which restricts its practical utility. In the case of ultrasound-based fetal biometry, an automated solution should ideally recognize key fetal structures in free-hand video guidance, select a standard plane from a video stream and perform biometry. A complete automated solution should automate all three sub-actions.Methods: This paper considers how to automate the complete human action of first-trimester biometry measurement from real-world freehand ultrasound. In the proposed hybrid Convolutional Neural Network (CNN) architecture design, a classification-regression-based guidance model detects and tracks fetal anatomical structures (using visual cues) in the ultrasound video. Several high-quality standard planes which contain the mid-sagittal view of the fetus are sampled at multiple timestamps (using a custom-designed confident-frame detector) based on the estimated probability values associated with predicted anatomical structures that define the biometry plane. Automated semantic segmentation is performed on the selected frames to extract fetal anatomical landmarks. A crown-rump length (CRL) estimate is calculated as the mean CRL from these multiple frames.Results: Our fully automated method shows a high correlation with clinical expert CRL measurement (Pearson ρ=0.92, R-Squared (R2)=0.84) and a low mean absolute error of 0.834 (weeks) for fetal age estimation on a test data set of 42 videos.Conclusion: A novel algorithm for standard plane detection employs a quality detection mechanism defined by clinical standards and ensuring precise biometric measurements.

Deep learning-based optical aberration estimation enables offline digital adaptive optics and super-resolution imaging

Optical aberrations degrade the performance of fluorescence microscopy. Conventional adaptive optics (AO) leverages specific devices, such as the Shack–Hartmann wavefront sensor and deformable mirror, to measure and correct optical aberrations. However, conventional AO requires either additional hardware or a more complicated imaging procedure, resulting in higher cost or a lower acquisition speed. In this study, we proposed a novel space-frequency encoding network (SFE-Net) that can directly estimate the aberrated point spread functions (PSFs) from biological images, enabling fast optical aberration estimation with high accuracy without engaging extra optics and image acquisition. We showed that with the estimated PSFs, the optical aberration can be computationally removed by the deconvolution algorithm. Furthermore, to fully exploit the benefits of SFE-Net, we incorporated the estimated PSF with neural network architecture design to devise an aberration-aware deep-learning super-resolution model, dubbed SFT-DFCAN. We demonstrated that the combination of SFE-Net and SFT-DFCAN enables instant digital AO and optical aberration-aware super-resolution reconstruction for live-cell imaging.

Automated machine learning and neural architecture optimization

Automated machine learning (AutoML) and neural architecture optimization (NAO) represent pivotal components in the landscape of machine learning and artificial intelligence. This paper extensively explores these domains, aiming to delineate their significance, methodologies, cutting-edge techniques, challenges, and emerging trends. AutoML streamlines and democratizes machine learning by automating intricate procedures, such as algorithm selection and hyperparameter tuning. Conversely, NAO automates the design of neural network architectures, a critical aspect for optimizing deep learning model performance. Both domains have made substantial advancements, significantly impacting research, industry practices, and societal applications. Through a series of experiments, classifier accuracy, NAO model selection based on hidden unit count, and learning curve analysis were investigated. The results underscored the efficacy of machine learning models, the substantial impact of architectural choices on test accuracy, and the significance of selecting an optimal number of training epochs for model convergence. These findings offer valuable insights into the potential and limitations of AutoML and NAO, emphasizing the transformative potential of automation and optimization within the machine learning field. Additionally, this study highlights the imperative for further research to explore synergies between AutoML and NAO, aiming to bridge the gap between model selection, architecture design, and hyperparameter tuning. Such endeavors hold promise in opening new frontiers in automated machine learning methodologies.

COMPARISON OF NEURAL NETWORK ARCHITECTURES TO PREDICTING ELECTRICITY CONSUMPTION BY ENTERPRISE

Forecasting of electricity consumption is a key tool for enterprises, energy supply and power grid organizations. Accurate forecasting enables to plan the distribution of limited resources of the power grid facilities, as well as to manage the demand for electricity. In the context of modern demand management technologies, improving the accuracy of forecasting of electricity consumption becomes especially important.
 
 The purpose of the study is to improve the accuracy of predicting power consumption by the power supply object using neural networks.
 
 Materials and methods. The work used a data set containing a profile of the enterprise’s capacity for a three-month period, as well as additional data, such as time of day, day of the week, weekends and holidays, month. The data set is divided into training and control parts. Preliminary data processing, neural network architecture design, training and testing were carried out. The criteria for the quality of training are the mean absolute error, the mean square error and the coefficient of determination.
 
 Research results. In the work, a comparative analysis of three neural network architectures was performed: a one-dimensional convolutional network, a recurrent neural network with long-term and short-term memory, and WaveNet, on the basis of which their indicators of the quality of power consumption forecasting were evaluated. It was shown that all considered architectures of neural networks are suitable for the use in the issue of predicting power consumption. Long-term and short-term memory networks have shown good results in power prediction due to their ability to handle long-term time dependencies. The WaveNet architecture outperformed both long-term and short-term memory model-based recurrent neural networks and one-dimensional convolutional networks by selected criteria.
 
 Conclusions. The study led to the conclusion that the use of neural networks, especially architectures with long-term and short-term memory and WaveNet, is an effective approach for predicting power consumption. The quality of forecasting significantly depends on the choice of hyperparameters and preliminary processing of the initial data. Prospect for further research in this area is studying the influence of various factors on the accuracy of forecasting and optimization of the learning process of neural networks.

Design of spiking neural network architecture based on dendritic computation principles

The paper presents the hardware architecture design of a spiking neural network (SNN) based on dendritic computation principles. The integration of active dendritic properties into the neuronal structure of SNN aims to minimize the number of functional blocks required for hardware implementation, including synaptic connections and neurons. The available memory on the neuromorphic architecture imposes limitations on implementation, hence the need to reduce the number of functional blocks. As a test task for the SNN based on dendritic computations, we selected the image classification of eight symbols, consisting of digits one through eight. These symbols are depicted as 3×7 pixel, 1-bit images. Active dendritic properties were analyzed using the “delay plasticity” [1] principle, which introduces the mechanism of adjusting input signal delays in spiking neuron inputs. We designed an SNN model with complementary delay inputs, referred to as the active dendrite SNN, as a principle implementation. Input spikes arriving at the primary inputs are duplicated to the delay inputs after a modifiable time delay. For convenience, each delay input was set at a single value. The input images were scanned sequentially. The neural network received three direct and three inverse inputs from the six main inputs that were coded with spikes corresponding to three pixels of a string. An “on” pixel was coded with a spike arriving at a direct input, while an “off” pixel was coded with a spike arriving at the corresponding inverse input. The line scanning time was 10 μs, input width was 1 μs, and delay time was 5 μs. The optimization of spiking neuron parameters was performed through a stochastic search algorithm based on simulated annealing. The parameters optimized for the Leaky-Integrate-and-Fire (LIF) neurons included the leakage time constant (22.8 μs), firing threshold (1150 arbitrary units), and refractory period (1 μs). The active dendrite SNN training employed the tempotron learning rule [2]. The training optimized the following parameters: the maximum change in synaptic weight on potentiation and depression (0.7 and –3 arbitrary units, respectively) and the synaptic weight’s upper bound (195 arbitrary units). Complementary delayed inputs facilitated the learning of the order in which input patterns arrived for SNN neurons during training. The paper compares an SNN architecture based on dendritic computations to our previously designed two-layer SNN with a hidden perceptron layer and an output layer consisting of LIF neurons [3]. Using the same LIF neuron design, input image coding, and LIF neuron layer structure as in the proposed architecture, our two-layer SNN with a hidden perceptron layer and output layer of LIF neurons successfully recognized 3×5 images of three symbols with only 10 neurons and 63 synapses. Alternatively, the active dendrite SNN was able to recognize 3×7 images of eight symbols with four neurons and 48 synaptic weights. In conclusion, incorporating active dendrite properties into the SNN architecture for image recognition resulted in optimized functional block usage, lowering the number of neurons and synapses used by 60 and 24%, respectively.

An enhanced multimodal fusion deep learning neural network for lung cancer classification

Cancer remains one of the leading causes of mortality worldwide, necessitating continuous advancements in early diagnosis and treatment. Deep learning, a subset of artificial intelligence, has emerged as a powerful tool in the field of medical image analysis, revolutionizing the way cancer is detected and diagnosed. The study discusses the various modalities employed in lung cancer diagnosis, such as medical imaging (e.g., radiology and pathology), genomics, and clinical data, highlighting the unique challenges associated with each domain. The proposed Multimodal Fusion Deep Neural Network (MFDNN) architecture design effectively integrates information from different modalities (e.g., medical imaging, genomics, clinical data) to enhance lung cancer diagnostic accuracy. Furthermore, it delves into the integration of clinical data, electronic health records, and multimodal approaches to improve the accuracy and reliability of lung cancer diagnosis. Moreover, we highlight the ethical considerations surrounding the deployment of Artificial Intelligence (AI) in clinical settings and the need for robust validation and regulatory guidelines. The Multimodal Fusion Deep Neural Network (MFDNN) achieves an exceptional accuracy rate of 92.5 %, marking a significant breakthrough in the realm of medical AI. MFDNN excels in precision, with 87.4 % accuracy in predicting cancer cases, and equally impresses in recall, capturing approximately 86.4 % of actual cancerous cases. The F1-score of 86.2 further exemplifies MFDNN's ability to strike a harmonious equilibrium, ensuring both diagnostic accuracy and minimized missed diagnoses. The performance is compared with established methods like CNN, DNN, and ResNet. The results underscore MFDNN's pivotal role in revolutionizing lung cancer diagnosis, promising more accurate and timely identification of this critical condition.

Evolutionary Architecture Optimization for Retinal Vessel Segmentation.

Retinal vessel segmentation (RVS) is crucial in medical image analysis as it helps identify and monitor retinal diseases. Deep learning approaches have shown promising results for RVS, but designing optimal neural network architecture is challenging and time-consuming. Neural architecture search (NAS) is a recent technique that automates the design of neural network architectures within a predefined search space. This study proposes a new NAS method for U-shaped networks, MedUNAS, that discovers deep neural networks with high segmentation performance and lower inference time for RVS problem. We perform opposition-based differential evolution (ODE) and genetic algorithm (GA) to search for the best network structure and compare discrete and continuous encoding strategies on the proposed search space. To the best of our knowledge, this is the first NAS study that performs ODE for RVS problems. The results show that the MedUNAS ODE and GA yield the best and second-best results regarding segmentation performance with less than 50% of the parameters of U-shaped state-of-the-art methods on most of the compared datasets. In addition, the proposed methods outperform the baseline U-Net on four datasets with networks with up to 15 times fewer parameters. Furthermore, ablation studies are performed to evaluate the generalizability of the generated networks to medical image segmentation problems that differ from the trained domain, revealing that such networks can be effectively adapted to new tasks with fine-tuning. The MedUNAS can be a valuable tool for automated and efficient RVS in clinical practice.

Learned Lossless Image Compression Through Interpolation With Low Complexity

With the increasing popularity of deep learning in image processing, many learned lossless image compression methods have been proposed recently. One group of algorithms are based on scale-based auto-regressive models and can provide competitive compression performance while also allowing easily parallelized computations and short encoding/decoding times. However, they use large neural networks and have high computational requirements. This paper presents an interpolation based learned lossless image compression method which falls in the scale-based auto-regressive models group. The method achieves compression performance better than or on par with the recent scale-based auto-regressive models, yet requires more than 10x less neural network parameters (0.19M) and encoding/decoding computation complexity. These achievements are due to the contributions/findings in the overall system and neural network architecture design, such as sharing interpolator neural networks across different scales, using separate neural networks for different parameters of the probability distribution model and performing the processing in the YCoCg-R color space instead of the RGB color space.

Hierarchical neural architecture search with adaptive global–local feature learning for Magnetic Resonance Image reconstruction

Neural architecture search (NAS) has been introduced into the design of deep neural network architectures for Magnetic Resonance Imaging (MRI) reconstruction since NAS-based methods can acquire the complex network architecture automatically without professional designing experience and improve the model’s generalization ability. However, current NAS-based MRI reconstruction methods suffer from a lack of efficient operators in the search space, which leads to challenges in effectively recovering high-frequency details. This limitation is primarily due to the prevalent use of convolution operators in the current search space, which struggle to capture both global and local features of MR images simultaneously, resulting in insufficient information utilization. To address this issue, a generative adversarial network (GAN) based model is proposed to reconstruct the MR image from under-sampled K-space data. Firstly, parameterized global and local feature learning modules at multiple scales are added into the search space to improve the capability of recovering high-frequency details. Secondly, to mitigate the increased search time caused by the augmented search space, a hierarchical NAS is designed to learn the global–local feature learning modules that enable the reconstruction network to learn global and local information of MR images at different scales adaptively. Thirdly, to reduce the number of network parameters and computational complexity, the standard operations in global–local feature learning modules are replaced with lightweight operations. Finally, experiments on several publicly available brain MRI image datasets evaluate the performance of the proposed method. Compared to the state-of-the-art MRI reconstruction methods, the proposed method yields better reconstruction results in terms of peak signal-to-noise ratio and structural similarity at a lower computational cost. Additionally, our reconstruction results are validated through a brain tumor classification task, affirming the practicability of the proposed method. Our code is available at https://github.com/wwHwo/HNASMRI.

EU-Net: Automatic U-Net neural architecture search with differential evolutionary algorithm for medical image segmentation

Medical images are crucial in clinical practice, providing essential information for patient assessment and treatment planning. However, manual extraction of information from images is both time-consuming and prone to errors. The emergence of U-Net addresses this challenge by automating the segmentation of anatomical structures and pathological lesions in medical images, thereby significantly enhancing the accuracy of image interpretation and diagnosis. However, the performance of U-Net largely depends on its encoder–decoder structure, which requires researchers with knowledge of neural network architecture design and an in-depth understanding of medical images. In this paper, we propose an automatic U-Net Neural Architecture Search (NAS) algorithm using the differential evolutionary (DE) algorithm, named EU-Net, to segment critical information in medical images to assist physicians in diagnosis. Specifically, by presenting the variable-length strategy, the proposed EU-Net algorithm can sufficiently and automatically search for the neural network architecture without expertise. Moreover, the utilization of crossover, mutation, and selection strategies of DE takes account of the trade-off between exploration and exploitation in the search space. Finally, in the encoding and decoding phases of the proposed algorithm, different block-based and layer-based structures are introduced for architectural optimization. The proposed EU-Net algorithm is validated on two widely used medical datasets, i.e., CHAOS and BUSI, for image segmentation tasks. Extensive experimental results show that the proposed EU-Net algorithm outperforms the chosen peer competitors in both two datasets. In particular, compared to the original U-Net, our proposed method improves the metric mIou by at least 6%.

Automatic Design of Convolutional Neural Network Architectures Under Resource Constraints.

With the rise of various smart electronics and mobile/edge devices, many existing high-accuracy convolutional neural network (CNN) models are difficult to be applied in practice due to the limited resources, such as memory capacity, power consumption, and spectral efficiency. In order to meet these constraints, researchers have carefully designed some lightweight networks. Meanwhile, to reduce the reliance on manual design on expert experience, some researchers also work to improve neural architecture search (NAS) algorithms to automatically design small networks, exploiting the multiobjective approaches that consider both accuracy and other important goals during optimization. However, simply searching for smaller network models is not consistent with the current research belief of "the deeper the better" and may affect the effectiveness of the model and thus waste the limited resources available. In this article, we propose an automatic method for designing CNNs architectures under constraint handling, which can search for optimal network models meeting the preset constraint. Specifically, an adaptive penalty algorithm is used for fitness evaluation, and a selective repair operation is developed for infeasible individuals to search for feasible CNN architectures. As a case study, we set the complexity (the number of parameters) as a resource constraint and perform multiple experiments on CIFAR-10 and CIFAR-100, to demonstrate the effectiveness of the proposed method. In addition, the proposed algorithm is compared with a state-of-the-art algorithm, NSGA-Net, and several manual-designed models. The experimental results show that the proposed algorithm can successfully solve the problem of the uncertain size of the optimal CNN model under the random search strategy, and the automatically designed CNN model can satisfy the predefined resource constraint while achieving better accuracy.

Dynamic Neural Network Architecture Design for Predicting Remaining Useful Life of Dynamic Processes

The prediction of remaining useful life is critical in predictive health management. This is done to reduce the expenses associated with operation and maintenance by avoiding errors and failures in dynamic processes. Recently, the abilities of feature classification and automated extraction of neural networks in its convolutional forms have shown fascinating performance when used for estimating the remaining useful life of dynamic processes using deep learning structures. This was accomplished by putting these talents to the task of predicting how long the procedures would be beneficial. Existing network topologies, on the other hand, virtually entirely extract features at a single scale while neglecting important information at other sizes. Meanwhile, because of the architecture of a single network path, the comprehensiveness of the features discovered by these tools is limited. To address these concerns, the authors propose a network structure based on a feature fusion strategy on a parallel multiscale architecture. This structure is then utilized to compute the remaining useful life. This prototype is divided into two sections: the first is a multiscale feature extraction module designed to extract local information features, and the second is a causal convolution module designed to extract global information features by combining multi-layer causal convolution with average pooling. The multiscale feature extraction module is intended for the extraction of local information features, while the causal convolution module is intended for the extraction of global information features. Finally, the two distinct paths are joined to create a fully integrated layer. The simulations and results show that this method has the potential to improve the efficiency and accuracy of estimating the remaining useful life index. Furthermore, the advantages of the established strategy are shown by comparing the results obtained with those produced by applying cutting-edge techniques on a well-known data-set depicting a simulated turbofan engine.

AutoInit: Analytic Signal-Preserving Weight Initialization for Neural Networks

Neural networks require careful weight initialization to prevent signals from exploding or vanishing. Existing initialization schemes solve this problem in specific cases by assuming that the network has a certain activation function or topology. It is difficult to derive such weight initialization strategies, and modern architectures therefore often use these same initialization schemes even though their assumptions do not hold. This paper introduces AutoInit, a weight initialization algorithm that automatically adapts to different neural network architectures. By analytically tracking the mean and variance of signals as they propagate through the network, AutoInit appropriately scales the weights at each layer to avoid exploding or vanishing signals. Experiments demonstrate that AutoInit improves performance of convolutional, residual, and transformer networks across a range of activation function, dropout, weight decay, learning rate, and normalizer settings, and does so more reliably than data-dependent initialization methods. This flexibility allows AutoInit to initialize models for everything from small tabular tasks to large datasets such as ImageNet. Such generality turns out particularly useful in neural architecture search and in activation function discovery. In these settings, AutoInit initializes each candidate appropriately, making performance evaluations more accurate. AutoInit thus serves as an automatic configuration tool that makes design of new neural network architectures more robust. The AutoInit package provides a wrapper around TensorFlow models and is available at https://github.com/cognizant-ai-labs/autoinit.

Online evolutionary neural architecture search for multivariate non-stationary time series forecasting

Time series forecasting (TSF) is one of the most important tasks in data science. TSF models are usually pre-trained with historical data and then applied on future unseen datapoints. However, real-world time series data is usually non-stationary and models trained offline usually face problems from data drift. Models trained and designed in an offline fashion can not quickly adapt to changes quickly or be deployed in real-time. To address these issues, this work presents the Online NeuroEvolution-based Neural Architecture Search (ONE-NAS) algorithm, which is a novel neural architecture search method capable of automatically designing and dynamically training recurrent neural networks (RNNs) for online forecasting tasks. Without any pre-training, ONE-NAS utilizes populations of RNNs that are continuously updated with new network structures and weights in response to new multivariate input data. ONE-NAS is tested on real-world, large-scale multivariate wind turbine data as well as the univariate Dow Jones Industrial Average (DJIA) dataset. Results demonstrate that ONE-NAS outperforms traditional statistical time series forecasting methods, including online linear regression, fixed long short-term memory (LSTM) and gated recurrent unit (GRU) models trained online, as well as state-of-the-art, online ARIMA strategies. Additionally, results show that utilizing multiple populations of RNNs which are periodically repopulated provide significant performance improvements, allowing this online neural network architecture design and training to be successful.

Automatic food bio-hazard detection system

This paper presents the design of a convolutional neural network architecture oriented to the detection of food waste, to generate a low, medium, or critical-level alarm. An architecture based on four convolution layers is used, for which a database of 100 samples is prepared. The database is used with the different hyperparameters that make up the final architecture, after the training process. By means of confusion matrix analysis, a 100% performance of the network is obtained, which delivers its output to a fuzzy system that, depending on the duration of the detection time, generates the different alarm levels associated with the risk.

A Neural Network Approach for Homogenization of Multiscale Problems

We propose a neural network-based approach to the homogenization of multiscale problems. The proposed method uses a derivative-free formulation of a training loss, which incorporates Brownian walkers to find the macroscopic description of a multiscale PDE solution. Compared with other network-based approaches for multiscale problems, the proposed method is free from the design of hand-crafted neural network architecture and the cell problem to calculate the homogenization coefficient. The exploration neighborhood of the Brownian walkers affects the overall learning trajectory. We determine the bounds of micro- and macro-time steps that capture the local heterogeneous and global homogeneous solution behaviors, respectively, through a neural network. The bounds imply that the computational cost of the proposed method is independent of the microscale periodic structure for the standard periodic problems. We validate the efficiency and robustness of the proposed method through a suite of linear and nonlinear multiscale problems with periodic and random field coefficients.

Thrust Prediction of Aircraft Engine Enabled by Fusing Domain Knowledge and Neural Network Model

Accurate prediction of aircraft engine thrust is crucial for engine health management (EHM), which seeks to improve the safety and reliability of aircraft propulsion. Thrust prediction is implemented using an on-board adaptive model for EHM. However, the conventional methods for building such a model are often tedious or overly data-dependent. To improve the accuracy of thrust prediction, domain knowledge can be leveraged. Hence, this study presents a strategy for building an on-board adaptive model that can predict aircraft engine thrust in real-time. The strategy combines engine knowledge and neural network architecture to construct a prediction model. The whole-model architecture is divided into separate modules that are mapped in a one-to-one form using a domain decomposition approach. The engine domain knowledge is used to guide feature selection and the neural network architecture design in the method. Furthermore, this study explains the relationships between aircraft engine features and how the model can predict engine thrust in flight condition. To demonstrate the effectiveness and robustness of the architecture, four different testing datasets were used for validation. The results show that the thrust prediction model created by the given architecture has maximum relative deviations below 4.0% and average relative deviations below 2.0% on all testing datasets. In comparison to the performance of the models created by conventional neural network architecture on the four testing datasets, the model created by the presented architecture proves more suitable for aircraft propulsion.

Using (1+1)D quantum cellular automata for exploring collective effects in large-scale quantum neural networks.

Central to the field of quantum machine learning is the design of quantum perceptrons and neural network architectures. A key question in this regard is the impact of quantum effects on the way such models process information. Here, we establish a connection between (1+1)D quantum cellular automata, which implement a discrete nonequilibrium quantum many-body dynamics through successive applications of local quantum gates, and quantum neural networks (QNNs), which process information by feeding it through perceptrons interconnecting adjacent layers. Exploiting this link, we construct a class of QNNs that are highly structured-aiding both interpretability and helping to avoid trainability issues in machine learning tasks-yet can be connected rigorously to continuous-time Lindblad dynamics. We further analyze the universal properties of an example case, identifying a change of critical behavior when quantum effects are varied, showing their potential impact on the collective dynamical behavior underlying information processing in large-scale QNNs.