Unified Neural Network Research Articles

Information and mathematical model of quantum communication channel state control processes

The most protected and stable communication systems today are quantum channels of information transmission and processing. Thanks to the unique properties of photons as information elements, it becomes possible to monitor and analyze the state of information flows in communication or information transmission channels. Physical attributes such as spin, polarization, radiation frequency, phase synchronization, and the quantum entanglement effect can be tracked and interpreted online to improve the quality and reliability of information in computer systems. In order to effectively use information in support or decision support systems, it is necessary to carefully formalize the processes and indicators of quantum systems for the creation of information processing and transmission, for which information and mathematical models should be created that describe the state of the quantum communication channel (QCC).The information model should allow the convolution of the information space. The mathematical model must prove the processes of tracking the states of quantum information and provide a description of the phase state of indicators of the quantum environment. A lock in a closed space with established cause-and-effect relationships is equal to a system of clear logic.The authors summarize the experience of developing and implementing the method of simulated dynamic modeling of events in an abstract communication channel, which allows formalizing and classifying cause-and-effect relationships of quantum carriers in the analyzed channels. It is proposed to use a unified neural network for the organization of SPPR in quantum-mechanical information transmission systems. Such a network could provide an automatic intelligent system state analysis mode. Such an analysis makes it possible to classify the aggregates of current system parameters to the level of diagnostics of the state of information flows and conclusions based on such diagnostics with the support of decision-making about the quality and reliability of the transmitted information. Such a system, working in OLAP (Online analytical processing) mode, could automatically manage the process of generation and transmission of information, reacting without human intervention to emerging critical errors or attempts at unauthorized system hacking. The observer effect leads to the fact that an attempt to measure the state of a photon inevitably causes an almost instantaneous change in this state. Attempting to parallelize a photon has the same consequences. This cannot be unnoticeable during further authorized acceptance of information. The analyzed quantum communication channel (QCM) consists of a set of technological elements distributed in space. The channel works in its own time, which is formed by clock pulses and creates a flow of information.

Applying vision-guided graph neural networks for adaptive task planning in dynamic human robot collaborative scenarios

ABSTRACT The Assemble-To-Order (ATO) strategy is increasingly becoming prevalent in the manufacturing sector due to the high demand for high-volume personalised and customised goods. The use of Human-Robot Collaborative (HRC) Systems are increasingly being investigated in order to make use of the dexterous strength of human hands while at the same time make use of the ability of robots to carry massive loads. However, current HRC systems struggle to adapt dynamically to varying human actions and cluttered workspaces. In this paper, we propose a novel neural network framework that integrates both Graph Neural Network (GNN) and Long Short-Term Memory (LSTM) for adaptive response during HRC scenarios. Our framework enables a robot to interpret human actions and generate detailed action plans while dealing with objects in a cluttered workspace thereby addressing the challenges of dynamic human-robot collaboration. Experimental results demonstrate improvements in assembly efficiency and flexibility, making our approach the first integration of iterative grasping and flexible HRC within a unified neural network architecture.

UNNIGSA: A Unified Neural Network Approach for Enhanced Stutter Detection and Gait Recognition Analysis

Stuttering, also known as stammering, is a speech disorder characterized by involuntary disruptions or disfluencies in a person's flow of speech. These disfluencies may include repetitions of sounds, syllables, or words; prolongations of sounds; and interruptions in speech known as blocks. This paper introduces Unified Neural Network for Integrated Gait and Speech Analysis (UNNIGSA), methodology that synergizes stutter detection (SD) and gait recognition through a unified neural network architecture. UNNIGSA is engineered to address two distinct yet interrelated challenges: the accurate detection of stuttering for enhanced beneficial interventions and the precise identification of individuals based on gait analysis. The system integrates a global attention mechanism to meticulously highlight salient features within speech patterns, thereby improving the accuracy of stutter classification and offering a potential leap forward in speech therapy practices. Additionally, UNNIGSA incorporates novel data processing techniques to manage the class imbalance prevalent in stuttering speech datasets, resulting in significantly enhanced performance over existing models. The methodology also extends the functionality of automatic speech recognition (ASR) systems, fostering greater inclusivity for individuals with speech disorders and enabling their more seamless interaction with virtual assistant technologies. Overall, UNNIGSA sets a new standard in the domains of speech disorder treatment and biometric identification, offering innovative solutions to long-standing challenges and paving the way for more inclusive and secure applications.

Neural networks for quantum state tomography with constrained measurements

Quantum state tomography (QST) aiming at reconstructing the density matrix of a quantum state plays an important role in various emerging quantum technologies. Recognizing the challenges posed by imperfect measurement data, we develop a unified neural network (NN)-based approach for QST under constrained measurement scenarios, including limited measurement copies, incomplete measurements, and noisy measurements. Through comprehensive comparison with other estimation methods, we demonstrate that our method improves the estimation accuracy in scenarios with limited measurement resources, showcasing notable robustness in noisy measurement settings. These findings highlight the capability of NNs to enhance QST with constrained measurements.

Toward Concurrent Identification of Human Activities with a Single Unifying Neural Network Classification: First Step.

The characterization of human behavior in real-world contexts is critical for developing a comprehensive model of human health. Recent technological advancements have enabled wearables and sensors to passively and unobtrusively record and presumably quantify human behavior. Better understanding human activities in unobtrusive and passive ways is an indispensable tool in understanding the relationship between behavioral determinants of health and diseases. Adult individuals (N = 60) emulated the behaviors of smoking, exercising, eating, and medication (pill) taking in a laboratory setting while equipped with smartwatches that captured accelerometer data. The collected data underwent expert annotation and was used to train a deep neural network integrating convolutional and long short-term memory architectures to effectively segment time series into discrete activities. An average macro-F1 score of at least 85.1 resulted from a rigorous leave-one-subject-out cross-validation procedure conducted across participants. The score indicates the method's high performance and potential for real-world applications, such as identifying health behaviors and informing strategies to influence health. Collectively, we demonstrated the potential of AI and its contributing role to healthcare during the early phases of diagnosis, prognosis, and/or intervention. From predictive analytics to personalized treatment plans, AI has the potential to assist healthcare professionals in making informed decisions, leading to more efficient and tailored patient care.

Significant phonon localization and suppressed transport in silicon-doped gallium oxide: A study using a unified neural network interatomic potential

Monoclinic gallium oxide (β-Ga2O3) is a fourth-generation semiconductor with great application potential in high-power microelectronics. Recent studies indicate that the electrical conductivity of β-Ga₂O₃ can be substantially enhanced through silicon (Si) doping. However, the effects on thermal transport, especially by considering the practical nanostructures within the crystal, have not yet been explored. To address this gap, we have developed a unified neural network potential for investigating the unexplored phonon transport of the β-(SixGa1–x)2O3 with varying doping levels. Our atomistic simulations showed that compared to intrinsic β-Ga2O3, the room-temperature thermal conductivities respectively decreased by 36.5%, 33.5%, and 38.8% along the a, b, and c axes in β-SiGa511O768, and by 79.6%, 74.9%, and 77.8% in β-SiGa7O12. The significant degradation in phonon transport is attributed to increased lattice anharmonicity, reduced sound velocity, and most importantly, induced phonon localization due to Si substitutions. A quantitative analysis reveals that the localization primarily occurs in phonons with frequencies exceeding 2.5 THz. The vibration is confined to a region around the Si atom, extending only to its second-nearest neighbors.

Integrated End-to-End automatic speech recognition for languages for agglutinative languages

The relevance of the problem of automatic speech recognition lies in the lack of research for low-resource languages, stemming from limited training data and the necessity for new technologies to enhance efficiency and performance. The purpose of this work was to study the main aspects of integrated end-to-end speech recognition and the use of modern technologies in the natural processing of agglutinative languages, including Kazakh. In this article, the study of language models was carried out using comparative, graphic, statistical and analytical-synthetic methods, which were used in combination. This paper addresses automatic speech recognition (ASR) in agglutinative languages, particularly Kazakh, through a unified neural network model that integrates both acoustic and language modeling. Employing advanced techniques like connectionist temporal classification and attention mechanisms, the study focuses on effective speech-to-text transcription for languages with complex morphologies. Transfer learning from high-resource languages helps mitigate data scarcity in languages such as Kazakh, Kyrgyz, Uzbek, Turkish, and Azerbaijani. The research assesses model performance, underscores ASR challenges, and proposes advancements for these languages. It includes a comparative analysis of phonetic and word-formation features in agglutinative Turkic languages, using statistical data. The findings aid further research in linguistics and technology for enhancing speech recognition and synthesis, contributing to voice identification and automation processes.

Improved multi-scale fusion network for solving non-smooth elliptic interface problems with applications

The utilization of deep learning methodologies for addressing partial differential equations (PDEs) has garnered significant attention in recent years. This paper introduces an improved network structure tailored for the discontinuity-capturing, enabling the resolution of interface problem through a unified neural network framework. Employing the probability space filling argument, we show that our model can generate convergent sequences, where the convergence rate depends on the number of sampling points. Several numerical experiments with regular and irregular interfaces are conducted to elucidate the convergence characteristics, thereby validating the theoretical assertions. Furthermore, we apply our approach to effectively solve the size-modified Poisson-Boltzmann test model, utilizing it for predicting electrostatics and the solvation free energies for proteins immersed in ionic solvents, thus showcasing practical applications of our method.

Detection of lanes, obstacles and drivable areas for self-driving cars using multifusion perception metrics

<p>Autonomous vehicles have been a recent trend and active research area from the onset of machine learning and deep learning algorithms. Computer vision and deep learning techniques have simplified the operations of continuous monitoring and decision-making capabilities of autonomous vehicles. A navigation system is facilitated by a visual system, where sensors and collectors process input in form of images or videos, and the navigation system will be making certain decisions to adhere to the safety of drivers and passers-by. This research article contemplates the model of obstacle detection, lane detection, and how the vehicle is supposed to act in terms of autonomous driving situation. This situation should resemble human driving conditions and should ensure maximum safety to both the stakeholders. A unified neural network for detecting lanes, objects, obstacles and to advise the driving speed is defined in this architecture. As far as autonomous driving is considered, these target elements are considered to be the predominant areas of focus for autonomous driving vehicles. Since capturing the images or videos have to be performed in real-time scenarios and processing them for relevant decision making have to be completed at a swift pace, a concept of context tensors is introduced in the decoders for discriminating the tasks based on priority. Every task is associated with the other tasks and also the decision-making process and hence this architecture will continue to learn every day. From the obtained results, it is evident that multitask networks can be improved using the proposed method in terms of accuracy, decision-making capability and reduced computational time. This model investigates the performance using Berkeley deep drive datasets which are considered to be a challenging dataset.</p>

CrossZoom: Simultaneous Motion Deblurring and Event Super-Resolving.

Even though the collaboration between traditional and neuromorphic event cameras brings prosperity to frame-event based vision applications, the performance is still confined by the resolution gap crossing two modalities in both spatial and temporal domains. This paper is devoted to bridging the gap by increasing the temporal resolution for images, i.e., motion deblurring, and the spatial resolution for events, i.e., event super-resolving, respectively. To this end, we introduce CrossZoom, a novel unified neural Network (CZ-Net) to jointly recover sharp latent sequences within the exposure period of a blurry input and the corresponding High-Resolution (HR) events. Specifically, we present a multi-scale blur-event fusion architecture that leverages the scale-variant properties and effectively fuses cross-modal information to achieve cross-enhancement. Attention-based adaptive enhancement and cross-interaction prediction modules are devised to alleviate the distortions inherent in Low-Resolution (LR) events and enhance the final results through the prior blur-event complementary information. Furthermore, we propose a new dataset containing HR sharp-blurry images and the corresponding HR-LR event streams to facilitate future research. Extensive qualitative and quantitative experiments on synthetic and real-world datasets demonstrate the effectiveness and robustness of the proposed method. Codes and datasets are released at https://bestrivenzc.github.io/CZ-Net/.

Vacancy-induced phonon localization in boron arsenide using a unified neural network interatomic potential

Vacancy-induced phonon localization in boron arsenide using a unified neural network interatomic potential

Statistical analysis of phased arrays with random excitation errors using a unified neural network architecture

AbstractRadiation patterns of a phased array tend to be affected by random excitation errors. In this paper, a unified neural network (UNN) architecture is presented to study the statistical characteristics of radiation patterns. Benefited from the inherent symmetry of the array manifolds and efficient data preprocessing, the UNN architectures are almost identical for planar arrays consisting of 4–10 000 elements except for a slight difference existed in the output layer. Therefore, this merit avoids repetitive selections of training hyperparameters and network architectures. This method can efficiently treat the problems with multiple observation points in order to obtain statistical radiation patterns, and an analytic or closed‐form solution is not required. Moreover, it can combine with the method of moments (MoM), which takes account of the element pattern and mutual coupling effects. The trained UNN models can predict the probability density function (PDF) of radiation characteristics at any spatial location. Numerical results show that the UNN and Monte Carlo (MC) results are comparable in accuracy.

Heterogeneous Domain Adaptation With Adversarial Neural Representation Learning: Experiments on E-Commerce and Cybersecurity.

Learning predictive models in new domains with scarce training data is a growing challenge in modern supervised learning scenarios. This incentivizes developing domain adaptation methods that leverage the knowledge in known domains (source) and adapt to new domains (target) with a different probability distribution. This becomes more challenging when the source and target domains are in heterogeneous feature spaces, known as heterogeneous domain adaptation (HDA). While most HDA methods utilize mathematical optimization to map source and target data to a common space, they suffer from low transferability. Neural representations have proven to be more transferable; however, they are mainly designed for homogeneous environments. Drawing on the theory of domain adaptation, we propose a novel framework, Heterogeneous Adversarial Neural Domain Adaptation (HANDA), to effectively maximize the transferability in heterogeneous environments. HANDA conducts feature and distribution alignment in a unified neural network architecture and achieves domain invariance through adversarial kernel learning. Three experiments were conducted to evaluate the performance against the state-of-the-art HDA methods on major image and text e-commerce benchmarks. HANDA shows statistically significant improvement in predictive performance. The practical utility of HANDA was shown in real-world dark web online markets. HANDA is an important step towards successful domain adaptation in e-commerce applications.

Global Spatiotemporal Graph Attention Network for Sea Surface Temperature Prediction

Accurately predicting sea surface temperature (SST) plays an important role in the study of marine ecosystems and global climate. The SST prediction problem is usually formulated as a time series regression problem, i.e., the future SST is predicted based on the historical SST. However, existing methods are typically devoted to modeling the highly nonlinear temporal correlations in SST data. They often ignore the dynamic spatial correlations that exist. This can limit the performance of these models, making accurately predicting SST challenging. To address this challenge, by combining graph neural networks that have a clear advantage in modeling spatial correlations, we propose a global spatio-temporal graph attention network (GSTGAT). Specifically, we capture the global dynamic spatial correlations of nodes through a global graph attention network module that fuses the static adjacency matrix learned adaptively by the graph learning module with the dynamic attention coefficients. A gated temporal convolutional network module is used to capture the nonlinear temporal correlations. Then, the above modules are integrated into a unified neural network to predict SST. We conduct experiments on multiple time-scale datasets in the Bohai Sea and the South China Sea. The experimental results show that GSTGAT achieves the best performance and consistently outperforms other methods for different prediction horizons in the two sea areas.

High-dimensional, slow-time-varying process monitoring technique based on adaptive eigen subspace extraction method

In this paper, in order to monitor the slow-time-varying industrial process, an adaptive method is proposed based on the neural network model and fault reconstruction method. Firstly, a unified neural network algorithm is introduced to extract the principal and minor eigen subspace with low computational complexity, and the whole eigenspace is divided into three partitions to further reduce the complexity of high-dimensional data computation. Then, the process is monitored based on a combined statistic index and the corresponding adaptive threshold. Moreover, the eigen subspace can still be updated even when in a faulty case. Finally, computer simulation confirms the capacity of the proposed method for high-dimensional, slow-time-varying process monitoring.

Flexible neural network for fast and accurate road scene perception

Accurate object detection on the road is the most important requirement of autonomous vehicles. Extensive work has been accomplished for car, pedestrian, and cyclist detection; however, comparatively, very few efforts have been put into 2D object detection. In this article, a dynamic approach is investigated to design a perfect unified neural network that could achieve the best results based on our available hardware. The proposed architecture is based on CSPNet for feature extraction in an end-to-end way. The net extracts visual features by using backbone subnet, visual object detection is based on a feature pyramid network (FPN). In order to increase the net flexibility, an auto-anchor generating method is applied to the detection layer that makes the net suitable for any datasets. For fine-tuning the net, activation, optimization, and loss functions are considered along with multiple check points. The proposed net is trained and tested based on the benchmark KITTI dataset. Our extensive experiments show that the proposed model for visual object detection is superior to others, where other nets output very low accuracy for pedestrian and cyclist detection, our proposed model achieves 99.3% recall rate based on our dataset.

Video Crowd Localization With Multifocus Gaussian Neighborhood Attention and a Large-Scale Benchmark.

Video crowd localization is a crucial yet challenging task, which aims to estimate exact locations of human heads in the given crowded videos. To model spatial-temporal dependencies of human mobility, we propose a multi-focus Gaussian neighborhood attention (GNA), which can effectively exploit long-range correspondences while maintaining the spatial topological structure of the input videos. In particular, our GNA can also capture the scale variation of human heads well using the equipped multi-focus mechanism. Based on the multi-focus GNA, we develop a unified neural network called GNANet to accurately locate head centers in video clips by fully aggregating spatial-temporal information via a scene modeling module and a context cross-attention module. Moreover, to facilitate future researches in this field, we introduce a large-scale crowd video benchmark named VSCrowd (https://github.com/HopLee6/VSCrowd), which consists of 60K+ frames captured in various surveillance scenes and 2M+ head annotations. Finally, we conduct extensive experiments on three datasets including our VSCrowd, and the experiment results show that the proposed method is capable to achieve state-of-the-art performance for both video crowd localization and counting.

A new target tracking filter based on deep learning

At present, current filters can basically solve the filtering problem in target tracking, but there are still many problems such as too many filtering variants, too many filtering forms, loosely coupled with the target motion model, and so on. To solve the above problems, we carry out cross-application research of artificial intelligence theory and methods in the field of tracking filters. We firstly analyze the computation graphs of typical α-β and Kalman. Through analysis, it is concluded that α-β and Kalman have the same computation structures analogous to a typical recurrent neural network and can be considered as a kind of recurrent neural network with constrained weights. Then, given this and considering that a recurrent neural network has the recognition capability for target motion patterns, a new filter is developed in a unified neural network architecture and specifically constructed using feedforward neural network, recurrent neural network, and attention mechanism. And the unified tracking filter proposed in this paper can generate three aspects of unity: a unified target motion model, an adaptive filter method, and an overall track filtering framework. Finally, Simulation results show that the proposed filter is effective and useful, of which the overall performance is superior to those of compared filters.

Dynamic Spatial-Temporal Representation Learning for Traffic Flow Prediction

As a crucial component in intelligent transportation systems, traffic flow prediction has recently attracted widespread research interest in the field of artificial intelligence (AI) with the increasing availability of massive traffic mobility data. Its key challenge lies in how to integrate diverse factors (such as temporal rules and spatial dependencies) to infer the evolution trend of traffic flow. To address this problem, we propose a unified neural network called Attentive Traffic Flow Machine (ATFM), which can effectively learn the spatial-temporal feature representations of traffic flow with an attention mechanism. In particular, our ATFM is composed of two progressive Convolutional Long Short-Term Memory (ConvLSTM <xref ref-type="bibr" rid="ref1" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">[1]</xref> ) units connected with a convolutional layer. Specifically, the first ConvLSTM unit takes normal traffic flow features as input and generates a hidden state at each time-step, which is further fed into the connected convolutional layer for spatial attention map inference. The second ConvLSTM unit aims at learning the dynamic spatial-temporal representations from the attentionally weighted traffic flow features. Further, we develop two deep learning frameworks based on ATFM to predict citywide short-term/long-term traffic flow by adaptively incorporating the sequential and periodic data as well as other external influences. Extensive experiments on two standard benchmarks well demonstrate the superiority of the proposed method for traffic flow prediction. Moreover, to verify the generalization of our method, we also apply the customized framework to forecast the passenger pickup/dropoff demands in traffic prediction and show its superior performance. Our code and data are available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/liulingbo918/ATFM</uri> .

Learned Hyperspectral Compression Using a Student’s T Hyperprior

Hyperspectral compression is one of the most common techniques in hyperspectral image processing. Most recent learned image compression methods have exhibited excellent rate-distortion performance for natural images, but they have not been fully explored for hyperspectral compression tasks. In this paper, we propose a trainable network architecture for hyperspectral compression tasks, which not only considers the anisotropic characteristic of hyperspectral images but also embeds an accurate entropy model using the non-Gaussian prior knowledge of hyperspectral images and nonlinear transform. Specifically, we first design a spatial-spectral block, involving a spatial net and a spectral net as the base components of the core autoencoder, which is more consistent with the anisotropic hyperspectral cubes than the existing compression methods based on deep learning. Then, we design a Student’s T hyperprior that merges the statistics of the latents and the side information concepts into a unified neural network to provide an accurate entropy model used for entropy coding. This not only remarkably enhances the flexibility of the entropy model by adjusting various values of the degree of freedom, but also leads to a superior rate-distortion performance. The results illustrate that the proposed compression scheme supersedes the Gaussian hyperprior universally for virtually all learned natural image codecs and the optimal linear transform coding methods for hyperspectral compression. Specifically, the proposed method provides a 1.51% to 59.95% average increase in peak signal-to-noise ratio, a 0.17% to 18.17% average increase in the structural similarity index metric and a 6.15% to 64.60% average reduction in spectral angle mapping over three public hyperspectral datasets compared to the Gaussian hyperprior and the optimal linear transform coding methods.