Neural Model Research Articles

The neural network algorithm-based quality assessment method for university English translation

ABSTRACT These results highlight the transformative potential of neural network algorithms in providing consistency and transparency while reducing the inherent subjectivity in human evaluations, revolutionizing translation quality assessment in academia. The findings have significant implications for academia, as reliable translation quality evaluations are crucial for fostering cross-cultural knowledge exchange. However, challenges such as domain-specific adaptation require further investigation to improve and maximize the effectiveness of this novel approach, ultimately enhancing the accessibility of academic content and promoting global academic discourse. The proposed method involves using neural network algorithms for assessing college-level English translation quality, starting with data collection and preparation, developing a neural network model, and evaluating its performance using human assessment as a benchmark. The study employed both human evaluators and a neural network model to assess the quality of translated academic papers, revealing a strong correlation (0.84) between human and model assessments. These findings suggest the model’s potential to enhance translation quality in academic settings, though additional research is needed to address certain limitations. The results show that the Neural Network-Based Model achieved higher scores in accuracy, precision, F-measure, and recall compared to Traditional Manual Evaluation and Partial Automated Model, indicating its superior performance in evaluating translation quality.

Low resource neural machine translation model optimization based on semantic confidence weighted alignment

Low resource neural machine translation model optimization based on semantic confidence weighted alignment

ASV station keeping under wind disturbances using neural network simulation error minimization model predictive control

AbstractStation keeping is an essential maneuver for autonomous surface vehicles (ASVs), mainly when used in confined spaces, to carry out surveys that require the ASV to keep its position or in collaboration with other vehicles where the relative position has an impact over the mission. However, this maneuver can become challenging for classic feedback controllers due to the need for an accurate model of the ASV dynamics and the environmental disturbances. This work proposes a model predictive controller using neural network simulation error minimization (NNSEM–MPC) to accurately predict the dynamics of the ASV under wind disturbances. The performance of the proposed scheme under wind disturbances is tested and compared against other controllers in simulation, using the robotics operating system and the multipurpose simulation environment Gazebo. A set of six tests was conducted by combining two varying wind speeds that are modeled as the Harris spectrum and three wind directions (, , and ). The simulation results clearly show the advantage of the NNSEM–MPC over the following methods: backstepping controller, sliding mode controller, simplified dynamics MPC (SD‐MPC), neural ordinary differential equation MPC (NODE‐MPC), and knowledge‐based NODE MPC. The proposed NNSEM–MPC approach performs better than the rest in five out of the six test conditions, and it is the second best in the remaining test case, reducing the mean position and heading error by at least % ( m) and % (), respectively, across all the test cases. In terms of execution speed, the proposed NNSEM–MPC is at least 36% faster than the rest of the MPC controllers. The field experiments on two different ASV platforms showed that ASVs can effectively keep the station utilizing the proposed method, with a position error as low as m and a heading error as low as within time windows of at least s. This would increase the potential applications of ASVs for launch, recovery, and replenishment in long‐term surveys in collaboration with other autonomous systems.

Pre-Trained Language Models for Text Generation: A Survey

Text Generation aims to produce plausible and readable text in human language from input data. The resurgence of deep learning has greatly advanced this field, in particular, with the help of neural generation models based on pre-trained language models (PLMs). Text generation based on PLMs is viewed as a promising approach in both academia and industry. In this article, we provide a survey on the utilization of PLMs in text generation. We begin with introducing two key aspects of applying PLMs to text generation: (1) how to design an effective PLM to serve as the generation model; and (2) how to effectively optimize PLMs given the reference text and to ensure that the generated texts satisfy special text properties. Then, we show the major challenges that have arisen in these aspects, as well as possible solutions for them. We also include a summary of various useful resources and typical text generation applications based on PLMs. Finally, we highlight the future research directions which will further improve these PLMs for text generation. This comprehensive survey is intended to help researchers interested in text generation problems to learn the core concepts, the main techniques and the latest developments in this area based on PLMs.

An effective ensemble approach for classification of chest X-ray images having symptoms of COVID: A precautionary measure for the COVID-19 subvariants

In recent years many people have died due to COVID-19 infection. The death rate increases due to the lack of proper diagnosis and treatment. It can be detected by RT-PCR test and from chest X-rays. Chest X-ray images can identify lung diseases such as COVID-19, viral pneumonia, and tuberculosis (TB). Because many illnesses share similar patterns and diagnoses, radiologists and clinicians find it challenging to distinguish between them. This study uses convolutional neural networks (CNN) and ensemble models (VGG-19, ResNet50, vision transformer) to analyze chest X-ray images from internet sources to identify lung illness. The present research proposed an ensemble model comprised of VGG-19, ResNet50, and a vision transformer for the detection of COVID-19 and normal images with an accuracy of 92.46 percent and 94.52 % for the detection of COVID-19, normal and viral pneumonia respectively. Also, a comparative study has been conducted between CNN and ensemble approach for detecting COVID-19 and pneumonia disease from chest x-ray images. The proposed approach may also help in the early detection of other variants of COVID-19 such as JN.1, and HV.1 in the future.

WindSeer: real-time volumetric wind prediction over complex terrain aboard a small uncrewed aerial vehicle

Real-time high-resolution wind predictions are beneficial for various applications including safe crewed and uncrewed aviation. Current weather models require too much compute and lack the necessary predictive capabilities as they are valid only at the scale of multiple kilometers and hours – much lower spatial and temporal resolutions than these applications require. Our work demonstrates the ability to predict low-altitude time-averaged wind fields in real time on limited-compute devices, from only sparse measurement data. We train a deep neural network-based model, WindSeer, using only synthetic data from computational fluid dynamics simulations and show that it can successfully predict real wind fields over terrain with known topography from just a few noisy and spatially clustered wind measurements. WindSeer can generate accurate predictions at different resolutions and domain sizes on previously unseen topography without retraining. We demonstrate that the model successfully predicts historical wind data collected by weather stations and wind measured by drones during flight.

A Simple Artificial Neural Model to Predict Dambovita River Temperature Affected by Urban Heat Islands in Bucharest City

Water bodies can offer local microclimates that have the potential to attenuate the effects of urban heat islands by reducing local temperature. This capability is shaded when the river is channelized. In such cases, the river temperature rises during hot periods, leading to negative impacts on the water quality. The main aim of this paper is to develop a local simple model to predict the temperature of the Dâmbovița River at its exit from Bucharest City, the capital of Romania. The location is chosen based on the historical critical impacts, in terms of extreme heatwaves that took place during hot summers, as well as future possible risks due to climate change. The water temperature prediction model is based on an artificial neural network that uses the Levenberg–Marquardt algorithm, due to its stability and rapid convergence capabilities. The model forecasts, with an accuracy of ±1 °C, the water temperature in an ungauged, downstream location, as a function of measured air and upstream water temperatures. The proposed model represents a first attempt to provide water managers in Bucharest City with a useful tool that will allow them to take timely measures to counteract the unwanted effects that can be generated by high water temperatures.

Recurrent neural network plasticity models: Unveiling their common core through multi-task learning

Recurrent neural network models are known to be particularly suitable for data-driven constitutive modeling due to their built-in memory variables. The main challenge preventing their widespread application to engineering materials lies in their excessive need of data for training. Here, we postulate that RNN models of elasto-plastic solids feature a large common core that is shared by all materials of the same class. The common core is complemented by material-specific layers with parameters that vary from material-to-material. After training RNN models for 25 different von Mises materials, we identify the common core of a multi-task model. Furthermore, we demonstrate through ensemble transfer learning that adding a new material to the multi-task model requires datasets that are two to three orders of magnitude smaller than the datasets needed for training an RNN from scratch. In addition, to multi-task learning, we introduce probabilistic ensembles of RNN plasticity models to quantify the epistemic uncertainty. A deep drawing simulation is performed to demonstrate the superior generalization capabilities of RNNs identified via multi-task learning as compared to those obtained through single task training.

Large language models for biomedicine: foundations, opportunities, challenges, and best practices.

Generative large language models (LLMs) are a subset of transformers-based neural network architecture models. LLMs have successfully leveraged a combination of an increased number of parameters, improvements in computational efficiency, and large pre-training datasets to perform a wide spectrum of natural language processing (NLP) tasks. Using a few examples (few-shot) or no examples (zero-shot) for prompt-tuning has enabled LLMs to achieve state-of-the-art performance in a broad range of NLP applications. This article by the American Medical Informatics Association (AMIA) NLP Working Group characterizes the opportunities, challenges, and best practices for our community to leverage and advance the integration of LLMs in downstream NLP applications effectively. This can be accomplished through a variety of approaches, including augmented prompting, instruction prompt tuning, and reinforcement learning from human feedback (RLHF). Our focus is on making LLMs accessible to the broader biomedical informatics community, including clinicians and researchers who may be unfamiliar with NLP. Additionally, NLP practitioners may gain insight from the described best practices. We focus on 3 broad categories of NLP tasks, namely natural language understanding, natural language inferencing, and natural language generation. We review the emerging trends in prompt tuning, instruction fine-tuning, and evaluation metrics used for LLMs while drawing attention to several issues that impact biomedical NLP applications, including falsehoods in generated text (confabulation/hallucinations), toxicity, and dataset contamination leading to overfitting. We also review potential approaches to address some of these current challenges in LLMs, such as chain of thought prompting, and the phenomena of emergent capabilities observed in LLMs that can be leveraged to address complex NLP challenge in biomedical applications.

Neural network assisted for parameters uncertainty of isogrid tube under different load types

Isogrid structures offer an excellent balance between mass and rigidity, making them a crucial component in various industrial sectors. However, like all types of structures, isogrids are susceptible to performance degradation due to uncertainties in geometrical and material parameters arising from manufacturing errors and exposure to critical service environments. This paper addresses the effects of parameter uncertainty on the structural behavior of isogrid tube structures. To this end, a Sobol index is employed to assess the sensitivity of the tube’s natural frequency, axial buckling, torsional buckling, and overall mass to three geometrical and one material parameter, each subject to a 10% random variation. The data for this analysis were generated using a parametrized numerical model specifically developed for this purpose. To optimize computational costs further, an artificial neural network predictive model is proposed, which accurately replicates the behavior of the numerical model. The results indicate that the angle between helicoidal ribs is the most influential parameter.

Bayesian data-driven framework for structural health monitoring of composite structures under limited experimental data

Ultrasonic-guided waves can be used to monitor the health of thin-walled structures. However, the run of experimental damage tests on materials like carbon fiber-reinforced plastics can be impractical and costly. Instead, numerical models can be used to create hybrid datasets to train machine learning algorithms, integrating data from numerical and experimental tests. This paper presents a Bayesian-driven framework to compensate for limited experimental data regarding Lamb wave propagation in composite plates. Using Bayesian inference, the framework updates a numerical finite element model, considering observed uncertainties by sampling posterior probability density functions for input parameters using Markov–Chain Monte Carlo simulations with the Metropolis-Hastings algorithm. A neural network surrogate model speeds-up these simulations, leading to a model that replicates the uncertain experimental setup. This model then generates data to augment true experimental data. Finally, a one-dimensional convolutional neural network is trained on a three different datasets to analyze Lamb wave signals and assess damage. Comparing training strategies shows the hybrid dataset augmented by samples generated by the updated FE model gives the most accurate damage size predictions.

Neural network prediction model for site response analysis based on the KiK-net database

Currently, the most widely used site response analysis methods include linear, equivalent linear and nonlinear approaches. However, the results of these analyses gradually deviate from the true values as the site softens or the earthquake intensity increases. To overcome this issue, this study employs a 1D Convolutional Neural Network model to simulate the relationships between the surface and downhole motions. To ensure comprehensive coverage of various sites and earthquake conditions, a careful selection of stations and seismic intensities resulted in 56 stations and 34,538 seismic records from KiK-net. Additionally, 30 seismic intensity measures and three site condition parameters were chosen. The convolutional operation was utilized to consider the correlations between the parameters, enabling the prediction of surface motion intensity measures. The performance of the neural network model was assessed based on residual analysis, and the model’s generalization ability was tested using samples that were not included in the training dataset. The results demonstrate that the neural network can effectively predict surface motion intensity measures. Furthermore, in comparison with the equivalent linear and nonlinear methods that do not rely on site-specific detailed soil data, the 1D Convolutional Neural Network exhibits superior performance and maintains high prediction accuracy even for soft soil and strong earthquake scenarios, confirming the enhanced predictive capability of the proposed model.

Intelligent vehicle trajectory tracking control based on physics-informed neural network dynamics model

In order to solve the accuracy problem of trajectory tracking control method based on data-driven model, an intelligent vehicle trajectory tracking control method based on physics-informed neural network (PINN) vehicle dynamics model is proposed. Aiming at the problem of poor interpretability of data-driven model, a vehicle dynamics model based on the PINN is established, and the physics-driven deep learning method is used instead of the data-driven deep learning method to obtain the dynamic characteristics of the intelligent vehicle, to benefit from both the physical-based method and the data-driven method. A sequential training method is also proposed to solve the coupling problem when training multiple PINNs simultaneously. The model takes the nonlinearity of the neural network model and physical interpretability into consideration compared to the standard neural network model. Then, based on the PINN vehicle dynamics model, a trajectory tracking controller based on the iterative linear quadratic regulator (ILQR) control algorithm is developed. The optimal control law is derived by optimizing the ILQR control algorithm to implement the intelligent vehicle’s precise and stable tracking for the desired trajectory. The Levenberg-Marquardt (LM) algorithm and line search technology are used and damping factor adjustment rules are set up to enhance the convergence performance of the ILQR control algorithm. In order to verify the effectiveness of the proposed method, the simulation is conducted under the condition of double lane change. The simulation results demonstrate that the proposed method can track the reference trajectory accurately under the limited conditions. Its control performance is much better than other algorithms.

Real‐Time Neural Materials using Block‐Compressed Features

AbstractNeural materials typically consist of a collection of neural features along with a decoder network. The main challenge in integrating such models in real‐time rendering pipelines lies in the large size required to store their features in GPU memory and the complexity of evaluating the network efficiently. We present a neural material model whose features and decoder are specifically designed to be used in real‐time rendering pipelines. Our framework leverages hardware‐based block compression (BC) texture formats to store the learned features and trains the model to output the material information continuously in space and scale. To achieve this, we organize the features in a block‐based manner and emulate BC6 decompression during training, making it possible to export them as regular BC6 textures. This structure allows us to use high resolution features while maintaining a low memory footprint. Consequently, this enhances our model's overall capability, enabling the use of a lightweight and simple decoder architecture that can be evaluated directly in a shader. Furthermore, since the learned features can be decoded continuously, it allows for random uv sampling and smooth transition between scales without needing any subsequent filtering. As a result, our neural material has a small memory footprint, can be decoded extremely fast adding a minimal computational overhead to the rendering pipeline.

An Overview of Demand Analysis and Forecasting Algorithms for the Flow of Checked Baggage among Departing Passengers

The research on baggage flow plays a pivotal role in achieving the efficient and intelligent allocation and scheduling of airport service resources, as well as serving as a fundamental element in determining the design, development, and process optimization of airport baggage handling systems. This paper examines baggage checked in by departing passengers at airports. The crrent state of the research on baggage flow demand is first reviewed and analyzed. Then, using examples of objective data, it is concluded that while there is a significant correlation between airport passenger flow and baggage flow, an increase in passenger flow does not necessarily result in a proportional increase in baggage flow. According to the existing research results on the influencing factors of baggage flow sorting and classification, the main influencing factors of baggage flow are divided into two categories: macro-influencing factors and micro-influencing factors. When studying the relationship between the economy and baggage flow, it is recommended to use a comprehensive analysis that includes multiple economic indicators, rather than relying solely on GDP. This paper provides a brief overview of prevalent transportation flow prediction methods, categorizing algorithmic models into three groups: based on mathematical and statistical models, intelligent algorithmic-based models, and combined algorithmic models utilizing artificial neural networks. The structures, strengths, and weaknesses of various transportation flow prediction algorithms are analyzed, as well as their application scenarios. The potential advantages of using artificial neural network-based combined prediction models for baggage flow forecasting are explained. It concludes with an outlook on research regarding the demand for baggage flow. This review may provide further research assistance to scholars in airport management and baggage handling system development.

StoryCraft AI: Exploring Generative Approaches to Story Narration through AI

In this project, we demonstrate a Storytelling AI system, which can generate short stories and complementary illustrated images with minimal input from the user. The system makes use of a text generation model, a text-to-image synthesis network, and a neural style transfer model. The final project is deployed into a website where users can build their stories. The field of AI has made significant changes in various domains, including Natural Language Processing and Generative models. One captivating application of these advancements is in the realm of storytelling. This project introduces a novel approach to story narration using a Generative AI model, specifically leveraging the GPT architecture. Traditional storytelling involves human creativity, imagination, and the ability to craft engaging narratives. However, the integration of AI into storytelling brings about new opportunities and challenges. In this project, we delve into the methodologies and techniques used to train a GPT-based model for generating coherent and captivating stories. In the end, we found that using computers for storytelling can be exciting, but we need to work together to ensure the stories are great and meaningful. Key Words: Artificial Intelligence (AI), Generative Adversarial Network (GAN), Generative Pre-trained Transformer (GPT)

A tensor basis neural network-based turbulence model for transonic axial compressor flows

Traditional turbulence models encounter limitations when simulating intricate flows within transonic axial compressors. In contrast, recent advancements in machine learning turbulence models have demonstrated enhanced potential in refining the precision of turbulence modeling. Notably, the tensor basis neural network (TBNN) methodology has successfully developed non-linear eddy viscosity turbulence models. These models possess the capability to capture the anisotropic characteristic of Reynolds stress. However, applying machine learning non-linear eddy viscosity models to aircraft turbomachinery remains relatively infrequent. Current research mainly focuses on linear eddy viscosity turbulence models based on the Boussinesq hypothesis that presupposes isotropy in Reynolds stress. In this work, we introduce a non-linear eddy viscosity turbulence model, denoted as the k-ω-SST-TBNN model based on the TBNN framework. This model has been employed to simulate the transonic axial compressor NASA Rotor 37. The TBNN is trained using large eddy simulation (LES) results datasets. The importance of input scalar features is analyzed using the random forest method, from which significant and low-degree variables are selected as inputs for the TBNN. Furthermore, this study proposes including the turbulent Mach number as one of the extra features, representing fluid compressibility, thereby extending the computational mass flow rate range of compressor. Additionally, this paper proposes a method involving the weighted average of Reynolds stress, combining the high-precision but less robust TBNN predictions with results based on the Boussinesq assumption to enhance the turbulence model's robustness. The trained k-ω-SST-TBNN model undergoes validation on Rotor 37, where it exhibits a marked improvement in the prediction of overall performance, tip-gap vortex, and radial distribution of flow parameters. The model also displays a commendable capacity for generalization.

Research on the Prediction of Infiltration Depth of Xiashu Loess Slopes Based on Particle Swarm Optimized Back Propagation (PSO-BP) Neural Network

The Xiashu loess exhibits expansion when in contact with water and contraction when water is lost, making it highly susceptible to the influence of rainfall. Therefore, it is essential to investigate the infiltration behavior of rainwater in Xiashu loess slopes under various conditions. The depth of infiltration in slopes directly affects the depth of landslide failure and serves as an important indicator for studying slope infiltration characteristics; only a handful of academics have delved into its study. This article is based on on-site rainfall experiments on Xiashu loess slopes, using three main factors, rainfall intensity, rainfall duration, and slope angle, as discrimination indicators for the infiltration depth of Xiashu loess slopes. The particle swarm optimization algorithm is employed to optimize the BP neural network and establish a PSO-BP neural network prediction model. The experimental data are accurately predicted and compared with the multivariate nonlinear regression model and traditional BP neural network models. The results demonstrate that the PSO-BP neural network model exhibits a better fit and higher prediction accuracy than the other two models. This model provides a novel approach for rapidly determining the infiltration depth of Xiashu loess slopes under different rainfall conditions. The results of this study lay the foundation for the prediction of the landslide damage depth and infiltration of Xiashu loess slopes.

HBCVTr: an end-to-end transformer with a deep neural network hybrid model for anti-HBV and HCV activity predictor from SMILES

Hepatitis B and C viruses (HBV and HCV) are significant causes of chronic liver diseases, with approximately 350 million infections globally. To accelerate the finding of effective treatment options, we introduce HBCVTr, a novel ligand-based drug design (LBDD) method for predicting the inhibitory activity of small molecules against HBV and HCV. HBCVTr employs a hybrid model consisting of double encoders of transformers and a deep neural network to learn the relationship between small molecules’ simplified molecular-input line-entry system (SMILES) and their antiviral activity against HBV or HCV. The prediction accuracy of HBCVTr has surpassed baseline machine learning models and existing methods, with R-squared values of 0.641 and 0.721 for the HBV and HCV test sets, respectively. The trained models were successfully applied to virtual screening against 10 million compounds within 240 h, leading to the discovery of the top novel inhibitor candidates, including IJN04 for HBV and IJN12 and IJN19 for HCV. Molecular docking and dynamics simulations identified IJN04, IJN12, and IJN19 target proteins as the HBV core antigen, HCV NS5B RNA-dependent RNA polymerase, and HCV NS3/4A serine protease, respectively. Overall, HBCVTr offers a new and rapid drug discovery and development screening method targeting HBV and HCV.

Learned Inference of Annual Ring Pattern of Solid Wood

AbstractWe propose a method for inferring the internal anisotropic volumetric texture of a given wood block from annotated photographs of its external surfaces. The global structure of the annual ring pattern is represented using a continuous spatial scalar field referred to as the growth time field (GTF). First, we train a generic neural model that can represent various GTFs using procedurally generated training data. Next, we fit the generic model to the GTF of a given wood block based on surface annotations. Finally, we convert the GTF to an annual ring field (ARF) revealing the layered pattern and apply neural style transfer to render orientation‐dependent small‐scale features and colors on a cut surface. We show rendered results of various physically cut real wood samples. Our method has physical and virtual applications such as cut‐preview before subtractive fabricating solid wood artifacts and simulating object breaking.