Unidirectional Network Research Articles

Unidirectional Communications in Secure IoT Systems—A Survey

The security of Internet of Things (IoT) systems has consistently been a challenge, particularly in the context of critical infrastructure. One particular approach not yet employed in this domain is the unidirectional communication paradigm. This survey presents an analysis of the most prevalent unidirectional communication solutions, namely, data diodes, network pumps, unidirectional gateways, and unidirectional protocols. The objective of the survey is to present an analysis of the unidirectional communication methods that meet the requirements of IoT security. These methods are classified according to their implementation and operational mode. The survey analyzes the unidirectional communication solutions based on their performance, the level of security offered, the cost-effectiveness, and their cost of implementation. Additionally, it includes an analysis of the existing off-the-shelf unidirectional communication implementations found in the industry. Furthermore, it identifies some of the most important current issues and development directions.

Using Machine Learning for Climate Modelling: Application of Neural Networks to a Slow-Fast Chaotic Dynamical System as a Case Study

This paper explores the capabilities of two types of recurrent neural networks, unidirectional and bidirectional long short-term memory networks, to build a surrogate model for a coupled fast–slow dynamic system and predicting its nonlinear chaotic behaviour. The dynamical system in question, comprising two versions of the classical Lorenz model with a small time-scale separation factor, is treated as an atmosphere–ocean research simulator. In numerical experiments, the number of hidden layers and the number of nodes in each hidden layer varied from 1 to 5 and from 16 to 256, respectively. The basic configuration of the surrogate model, determined experimentally, has three hidden layers, each comprising between 16 and 128 nodes. The findings revealed the advantages of bidirectional neural networks over unidirectional ones in terms of forecasting accuracy. As the forecast horizon increases, the accuracy of forecasts deteriorates, which was quite expected, primarily due to the chaotic behaviour of the fast subsystem. All other things being equal, increasing the number of neurons in hidden layers facilitates the improvement of forecast accuracy. The obtained results indicate that the quality of short-term forecasts with a lead time of up to 0.75 model time units (MTU) improves most significantly. The predictability limit of the fast subsystem (“atmosphere”) is somewhat greater than the Lyapunov time.

Polarity management of multicore fiber-based optical devices in unidirectional and bidirectional spatial channel networks

Uncoupled multicore fibers (MCFs) are expected to be the first to be commercially deployed due to their high compatibility with existing single-mode fiber technologies. Since MCFs have a 3D shape, they generally exhibit connection polarity. Thus, optical devices based on MCFs also generally have polarity, which will complicate the core resource assignment and end-to-end core connections in future MCF-based spatial channel networks (SCNs). In this paper, we first discuss the polarity of MCF-based optical devices (MODs) such as MCF patch cords, spatial multiplexers (SMUXs), core selective switches (CSSs), and core selectors (CSs). We then propose a definition for global core numbers in a two-MCF unidirectional (2MCF-UD) SCN and a single-MCF bidirectional (1MCF-BD) SCN. We also propose a method for managing the polarity of MODs and correctly connecting cores end-to-end. To verify the effectiveness of the proposed global core numbering and polarity management method for MODs, testbeds emulating a 2MCF-UD SCN and a 1MCF-BD SCN are constructed using prototype CSS, CS, and SMUX devices. By using light with different optical frequencies as input and observing the output spectrum, we confirm that the spatial channel specified by the global core number is established correctly end-to-end in the SCN if the polarity of the MODs is set correctly.

НЕЙРОСЕТЕВОЙ ПОДХОД К КОГНИТИВНОМУ МОДЕЛИРОВАНИЮ

Some features of cognitive modeling are presented, including the prerequisites for a cognitive approach to solving complex problems. Cognitive modeling involves the use of various artificial neural networks, including convolutional neural networks. The classification of artificial neural networks according to various characteristics is given. The features of self-organizing neural networks and networks using deep learning methods are considered. The artificial neural network, which is a three-layer unidirectional direct propagation network, the interface of a computer program used to approximate functions using the specified neural network, as well as the solution of the image recognition problem using an artificial convolutional neural network, in which the neural network parameters are adjusted for each recognizable image fragment in order to adaptively filter the image, are considered in detail. The analysis of images in video surveillance systems in order to detect fires allows them to be detected at an early stage and, thus, prevent the fire propogation.

Coupled unidirectional chaotic microwave graphs.

We experimentally investigate the undirected open microwave network Γ with internal absorption composed of two coupled directed halves, unidirectional networks Γ_{+} and Γ_{-}, corresponding to two possible directions of motion on their edges. The two-port scattering matrix of the network Γ is measured and the spectral statistics and the elastic enhancement factor of the network are evaluated. The comparison of the number of experimental resonances with the theoretical one predicted by the Weyl's law shows that within the experimental resolution the resonances are doubly degenerate. This conclusion was also corroborated by the numerical calculations. Though the network is characterized by the time-reversal symmetry, the missing level spectral statistics and the elastic enhancement factor are rather close to the Gaussian unitary ensemble predictions in random matrix theory. We used numerical calculations for the open nondissipative quantum graph possessing the same structure as the microwave network Γ to investigate the doublet structures in the spectrum which otherwise would not be experimentally resolved. We show that the doublet size distribution is close to the Poisson distribution.

Harnessing the Missing Spectral Correlation for Metasurface Inverse Design.

A long-held tenet in computer science asserts that the training of deep learning is analogous to an alchemical furnace, and its "black box" signature brings forth inexplicability. For electromagnetic metasurfaces, the related intelligent applications also get stuck into such a dilemma. Although the past 5 years have witnessed a proliferation of deep learning-based works across complex photonic scenarios, they neglect the already existing but untapped physical laws. Here, the intrinsic correlation between the real and imaginary parts of the spectra are revealed using Kramers-Kronig relations, which is then mimicked by bidirectional information flow in neural network space. Such consideration harnesses the missing spectral connection to extract crucial features effectively. The bidirectional recurrent neural network is benchmarked in metasurface inverse design and compare it with a fully-connected neural network, unidirectional recurrent neural network, and attention-based transformer. Beyond the improved accuracy, the study examines the intermediate information products and physically explains why different network structures yield different performances. The work offers explicable perspectives to utilize physical information in the deep learning field and facilitates many data-intensive research endeavors.

A BlockChain-Based IoT Data Securing and Sharing System for Attendance Management

Objective: An Internet of Things (IoT) network has constraints on power consumption, processing power and security of end devices in the network. On the other hand, blockchain technology has some difficulties in storage capacity, data privacy, the ability to revoke consent, energy usage, scalability, and speed. To overcome the individual limitations of these two technologies, we integrate these two to propose a new design with an intention to authenticate, secure and share data pertaining to attendance system. In this present work the primary implementation details of our proposed work regarding coding are analyzed and a straightforward solution is offered by using indexing. Methods: Through our proposed work we introduce a multifaceted authentication system for user device substantiation, a Public Key Infrastructure (PKI) and RSA (Rivest-Shamir-Adleman) based blockchain implementation for data sharing and a rule-based access control scheme used for accessing data. Findings: The system can work with unidirectional and bidirectional IoT networks, and the data is stored on blockchain, providing more security to the data generated on the IoT network. Blockchain technology is utilized to offer a solution for certain problems encountered within a widespread network. Furthermore, a multi-factor authentication scheme is introduced and a rule-based access control scheme to access the data. Novelty: The blockchain is emerging which can be used to decentralize management and protect sensitive data. In a blockchain based attendance system, no administrative authority is permitted to modify or erase data. The individual who adds a data entry to the blockchain cannot later deny their involvement in the action. On the blockchain, all of the data and history are accessible to every participant. The blockchain based attendance system needs to offer a database that can be trusted, secure, and impossible to manipulate. Keywords: Blockchain (BC), Internet of Things (IoT), Student Attendance Management (SAM), User Authentication Key (UAK), Data Sharing

An optoacoustic field-programmable perceptron for recurrent neural networks.

Recurrent neural networks (RNNs) can process contextual information such as time series signals and language. But their tracking of internal states is a limiting factor, motivating research on analog implementations in photonics. While photonic unidirectional feedforward neural networks (NNs) have demonstrated big leaps, bi-directional optical RNNs present a challenge: the need for a short-term memory that (i) programmable and coherently computes optical inputs, (ii) minimizes added noise, and (iii) allows scalability. Here, we experimentally demonstrate an optoacoustic recurrent operator (OREO) which meets (i, ii, iii). OREO contextualizes the information of an optical pulse sequence via acoustic waves. The acoustic waves link different optical pulses, capturing their information and using it to manipulate subsequent operations. OREO's all-optical control on a pulse-by-pulse basis offers simple reconfigurability and is used to implement a recurrent drop-out and pattern recognition of 27 optical pulse patterns. Finally, we introduce OREO as bi-directional perceptron for new classes of optical NNs.

An efficient tool for Parkinson's disease detection and severity grading based on time-frequency and fuzzy features of cumulative gait signals through improved LSTM networks

Parkinson's disease (PD) is a widespread neurodegenerative condition that affects many individuals annually. Early identification and monitoring of disease progression are crucial to effectively managing symptoms and preventing motor complications. This research proposes an automated PD diagnosis and severity-grading model based on time-frequency and fuzzy features using improved uni-directional and bi-directional long short-term memory networks with sensitive hyperparameters optimization. We utilize vertical ground reaction force signals collected from Physionet's publicly available dataset recorded during regular and dual-task clinical trials of walking measurements. Only the cumulative signal of both feet was then utilized and segmented into 30-s windows without further pre-processing. Subsequently, we extracted only four key time-frequency and fuzzy features from each segment, effectively capturing the signal's inherent uncertainty. Bayesian optimization is employed in both detection and grading approaches to fine-tune the two critical hyperparameters: the initial learning rate and the number of hidden units in the network. The detection phase yields an exceptional accuracy of 99.19%, surpassing state-of-the-art studies with the same dataset. In the grading phase, classification based on the unified PD rating scale values achieves an accuracy of 92.28%. The proposed study delves into the potential of cumulative gait signals as a powerful diagnostic tool for PD, aiming to extract precise and intricate information by implementing straightforward and minimal processing endeavors. This method demonstrates significant efficiency in terms of complexity, cost, and energy consumption by utilizing a single-dimensional signal, eliminating the need for pre-processing steps, and limiting the features used for training.

Long short-term memory (LSTM)-based news classification model.

In this study, we used unidirectional and bidirectional long short-term memory (LSTM) deep learning networks for Chinese news classification and characterized the effects of contextual information on text classification, achieving a high level of accuracy. A Chinese glossary was created using jieba-a word segmentation tool-stop-word removal, and word frequency analysis. Next, word2vec was used to map the processed words into word vectors, creating a convenient lookup table for word vectors that could be used as feature inputs for the LSTM model. A bidirectional LSTM (BiLSTM) network was used for feature extraction from word vectors to facilitate the transfer of information in both the backward and forward directions to the hidden layer. Subsequently, an LSTM network was used to perform feature integration on all the outputs of the BiLSTM network, with the output from the last layer of the LSTM being treated as the mapping of the text into a feature vector. The output feature vectors were then connected to a fully connected layer to construct a feature classifier using the integrated features, finally classifying the news articles. The hyperparameters of the model were optimized based on the loss between the true and predicted values using the adaptive moment estimation (Adam) optimizer. Additionally, multiple dropout layers were added to the model to reduce overfitting. As text classification models for Chinese news articles, the Bi-LSTM and unidirectional LSTM models obtained f1-scores of 94.15% and 93.16%, respectively, with the former outperforming the latter in terms of feature extraction.

Development of Drainage Tools for Facilitating the Creation of Topologically-Sound, Unidirectional Flow Tile Networks

HighlightsNew publicly accessible drainage design QGIS plugins were developed.These tools were successfully used to create topologically-sound subsurface drainage layout networks.The tools are easy to use and reduce the level of expertise required to prepare drainage networks that can be successfully sized and buried.Abstract. An important step in designing a drainage system is laying out the network. Successful installation and sizing of the system require that the network be topologically sound and unidirectional. Online digitization of drainage systems does not always result in networks that meet these criteria. Tile lines may not be consistently digitized from upstream to downstream, and there may be disconnected segments, dangles, and overshoots. We have developed a set of open-source tools (plugins) in QGIS that convert a series of laterals and mains digitized on a map into a topologically-sound drainage network to facilitate both burying and sizing the drainage lines. Steps include reversing node numbering, if necessary, to ensure that tile nodes are numbered from upstream to downstream, connecting the ends of tile lines that are within a certain tolerance of each other, correcting overshoots, and setting an order for burying or calculating cumulative flow in line segments. An application example is presented using a proposed tile network layout at the university farm to demonstrate the capacity of these new tools to fix topologically disconnected tile lines and generate sound tile networks that can be successfully buried and sized. These easy-to-use tools are compatible with all versions of QGIS3 and may be freely downloaded or installed from the official QGIS Plugin Repository. Keywords: Line Geometry, Plugins, QGIS Software, Tile Layout Routines, Tile Nodes and Lines.

EMG-based Simultaneous Estimations of Joint Angle and Torque during Hand Interactions with Environments.

It is necessary to control contact force through modulation of joint stiffness in addition to the position of our limb when manipulating an object. This is achieved by contracting the agonist muscles in an appropriate magnitude, as well as, balancing it with contraction of the antagonist muscles. Here we develop a decoding technique that estimates both the position and torque of a joint of the limb in interaction with an environment based on activities of the agonist-antagonistic muscle pairs using electromyography in real time. The long short-term memory (LSTM) network that is capable of learning time series of a longtime span with varying time lags is employed as the core processor of the proposed technique. We tested both the unidirectional LSTM network and bidirectional LSTM network. A validation was conducted on the wrist joint moving along a given trajectory under resistance generated by a robot. The decoding approach provided an agreement of greater than 93% in kinetics (i.e. torque) estimation and an agreement of greater than 83% in kinematics (i.e. angle) estimation, between the actual and estimated variables, during interactions with an environment. We found no significant differences in performance between the unidirectional LSTM and bidirectional LSTM as the learning device of the proposed decoding method.

The evolution of cooperation in the unidirectional division of labour on a tree network.

Division of labour on complex networks is rarely investigated using evolutionary game theory. We investigate a division of labour where divided roles are assigned to groups on the nodes of a general unidirectional finite tree graph network. From the network's original node, a task flows and is divided along the branches. A player is randomly selected in each group of cooperators and defectors, who receives a benefit from a cooperator in the upstream group and a part of the task. A cooperator completes their part by paying a cost and then passing it downstream until the entire task is completed. Defectors do not do anything and the division of labour stops, causing all groups to suffer losses due to the incomplete task. We develop a novel method to analyse the local stability in this general tree. We discover that not the benefits but the costs of the cooperation influence the evolution of cooperation, and defections in groups that are directly related to that group's task cause damage to players in that group. We introduce two sanction systems, one of which induces the evolution of cooperation more than the system without sanctions, and promote the coexistence of cooperator and defector groups.

Feature Importance Ranking of Random Forest-Based End-to-End Learning Algorithm

Efficient land management and farming practices are critical to maintaining agricultural production, especially in Europe with limited arable land. It is very time consuming to rely on a manual field inspection of cultivated land to archive farm crops. But with the help of satellite monitoring data on the earth’s surface, it is a new vision to classify farmland based on deep learning. This article has studied the Sentinel 2 (S2) data, which are top-of-atmosphere (TOA) reflectance values at the processing level-1C (L1C) observed from some areas of Germany and France. Aiming at the problem that the interference of atmosphere and cloud coverage weakens the recognition accuracy of subsequent algorithms, a method of combining feature expansion and feature importance analysis is proposed to optimize the raw S2 data. Specifically, the new 13 spectral features are expanded based on the linear and nonlinear combination of the raw 13 spectral bands of S2. The random forest (RF) algorithm is used to score the importance of features, and the important features of each time series are selected to form a new dataset. Then, an end-to-end deep learning model has been used for training. The structure of the model is a two-layer unidirectional recurrent neural network with long short-term memory (LSTM) as the backbone. And two linear layers as the output, which form two decision-making heads, respectively, representing output classification probability and the stop decision. The results show that adding features and selecting features is beneficial for the model to improve classification accuracy and predict the classification without all of the input data. This end-to-end classification pattern with early prediction would support intelligent monitoring of farm crops with a great advantage to the implementation of various agricultural policies.

Unidirectional Ring-Based WDM Fiber Network for Both Downlink and Uplink Signal Access

In the paper, a dual-bidirectional ring-type wavelength-division-multiplexing (WDM) access network with downlink and uplink signal access simultaneously using a single fiber backbone in clockwise and counterclockwise directions, respectively. The proposed network architecture is simple and easy to implement via the designed remote node (RN) and optical line termination (OLT) modules, but it also can double the downlink traffic using the original WDM downlink wavelengths. The presented ring-type WDM network can also avoid the Rayleigh backscattering (RB) beat noise when the same wavelengths are applied as downlink and uplink channels concurrently. In the measurement, 50 km long-reach and 15 km short-reach fiber transmission lengths are achieved for the symmetrical 10 and 28 Gbit/s on-off keying (OOK) data access, respectively. In addition, based on the obtained power budgets of eight downlink WDM signals and network design at the forward error correction (FEC) threshold, 16 optical network units (ONUs) can be supported simultaneously.

Interdependent security games in a unidirectional network.

We consider directed tree networks with a single source, where there exists a positive probability of a disruptive event at any node. Such networks model security considerations in pipelines as well as in unidirectional digital networks. If a disruptive event occurs at a certain node, that node and its downstream nodes incur economic losses. Users thus have an incentive to invest in upstream locations as well as their own sites to reduce the probability of a disruptive event. The initial model we develop to reduce the expected investment plus disruption costs is a multiplicative model for which closed-form solutions cannot be obtained in general. We overcome this problem with an additive model that we show closely approximates the initial formulation. This model reduces the security problem to a public goods setting where we minimize the total expected cost at each node. The users then need to share these costs in an equitable fashion, which gives rise to a set of cooperative games. For the case where disutilities to all users are identical, the Shapley value can be computed efficiently, along the lines of an Airport Game. We also treat the case where risk reduction and disutility vary across the network. Finally, we prove that the cooperative game is concave in this general case, which guarantees that the core of the game is nonempty and that the Shapley value is an element of the core.

Synapses do not facilitate prion-like transfer of alpha-synuclein: a quantitative study in reconstructed unidirectional neural networks

Alpha-synuclein (aSyn) aggregation spreads between cells and underlies the progression of neuronal lesions in the brain of patients with synucleinopathies such as Parkinson’s diseases. The mechanisms of cell-to-cell propagation of aggregates, which dictate how aggregation progresses at the network level, remain poorly understood. Notably, while prion and prion-like spreading is often simplistically envisioned as a “domino-like” spreading scenario where connected neurons sequentially propagate protein aggregation to each other, the reality is likely to be more nuanced. Here, we demonstrate that the spreading of preformed aSyn aggregates is a limited process that occurs through molecular sieving of large aSyn seeds. We further show that this process is not facilitated by synaptic connections. This was achieved through the development and characterization of a new microfluidic platform that allows reconstruction of binary fully oriented neuronal networks in vitro with no unwanted backward connections, and through the careful quantification of fluorescent aSyn aggregates spreading between neurons. While this allowed us for the first time to extract quantitative data of protein seeds dissemination along neural pathways, our data suggest that prion-like dissemination of proteinopathic seeding aggregates occurs very progressively and leads to highly compartmentalized pattern of protein seeding in neural networks.

Unidirectional Neuromorphic Resistive Memory Integrated with Piezoelectric Nanogenerator for Self‐Power Electronics

AbstractThis study presents a method to enhance data processing by integrating a unidirectional analogue artificial neuromorphic memristor device with a piezoelectric nanogenerator, taking inspiration from biological information processing. A self‐powered unidirectional neuromorphic resistive memory device is proposed, comprising an ITO/ZnO/Yb2O3/Au structure combined with a high‐sensitivity piezoelectric nanogenerator (PENG) ITO/ZnO/Al. The memristor device is operated at a voltage sweep of ±4 V with a low operating current in a range of 1.4 µA. The filament formation is studied using a conductive mode atomic force microscope. The integration enables the creation of a self‐powered artificial sensing system that converts mechanical stimuli from the PENG into electrical signals, which are subsequently processed by analogue unidirectional neuromorphic device to mimic the functionality of a neuron without requiring additional circuitry. This emulation encompasses crucial functions such as potentiation, depression, and synaptic plasticity. Furthermore, this study highlights the potential for hardware implementations of neural networks with a weight change of memristor device with nonlinearity (NL) of potentiation and depression of 1.94 and 0.89, respectively, with an accuracy of 93%. The outcomes of this research contribute to the progress of next‐generation low‐power, self‐powered unidirectional neuromorphic perception networks with correlated learning and trainable memory capabilities.

BiTCN_DRSN: An effective software vulnerability detection model based on an improved temporal convolutional network

The detection of software vulnerabilities is a challenging task in the field of security. With the increasing scale of software and the rapid development of artificial intelligence technology, deep learning has been extensively applied to automatic vulnerability detection. Temporal Convolutional Networks (TCNs) have been shown to perform well in tasks that can be processed in parallel; they can adaptively learn complex structures (including in-time series data); and they have exhibited stable gradients — they are relatively easier to train, and can quickly converge to an optimal solution. However, TCNs cannot simultaneously capture the bidirectional semantics of the source code, since they do not have a bidirectional network structure. Furthermore, because of the weak noise resistance of residual TCN connections, TCNs are also susceptible to learning features that are not related to vulnerabilities when learning the source code features. To overcome the limitations of the traditional TCN, we propose a bidirectional TCN model based on the Deep Residual Shrinkage Network (DRSN), namely BiTCN_DRSN. BiTCN_DRSN combines TCN and DRSN to enhance the noise immunity and make the network model more attentive to the features associated with vulnerabilities. In addition, addressing the limitation that the TCN is a unidirectional network structure, the forward and backward sequences are utilized for bidirectional source-code feature learning. The experimental results show that the proposed BiTCN_DRSN model can effectively improve the accuracy of source-code vulnerability detection, compared with some existing neural-network models. Compared with the traditional TCN, our model increases the accuracy by 4.22%, 2.42% and 2.66% on the BE-ALL, RM-ALL and HY-ALL datasets, respectively. The proposed BiTCN_DRSN model also exhibits improved detection stability.

Tree-Based Unidirectional Neural Networks for Low-Power Computer Vision

This article describes the novel Tree-based Unidirectional Neural Network (TRUNK) architecture. This architecture improves computer vision efficiency by using a hierarchy of multiple shallow Convolutional Neural Networks (CNNs), instead of a single very deep CNN. We demonstrate this architecture’s versatility in performing different computer vision tasks efficiently on embedded devices. Across various computer vision tasks, the TRUNK architecture consumes 65% less energy and requires 50% less memory than representative low-power CNN architectures, e.g., MobileNet v2, when deployed on the NVIDIA Jetson Nano.