Output Nodes Research Articles

Analysis of the Use of Artificial Neural Network Models in Predicting Bitcoin Prices

Bitcoin is one of the fastest-growing digital currencies or cryptocurrencies in the world. However, the highly volatile Bitcoin price poses a very extreme risk for traders investing in cryptocurrencies, especially Bitcoin. To anticipate these risks, a prediction system is needed to predict the fluctuations in cryptocurrency prices. Artificial Neural Network (ANN) is a relatively new model discovered and can solve many complex problems because the way it works mimics human nerve cells. ANN has the advantage of being able to describe both linear and non-linear models with a fairly wide range. This research aims to determine the best performance and level of accuracy of the ANN model using the Back-Propagation Neural Network (BPNN) algorithm in predicting Bitcoin prices. This study uses Bitcoin price data for the period 2020 to 2023 taken from the CoinDesk market. The results of this study indicate that the ANN model produces the best performance in the form of four input nodes, 12 hidden nodes, and one output node (4-12-1) with an accuracy rate of around 3.0617175%.

1-Norm twin random vector functional link networks based on Universum data for leaf disease detection

Due to rapid climate change and man-made activities, the types of leaf diseases are gradually increasing. As a result, taking the essential measures to recognize and diagnose the problem becomes a top priority. Though manual techniques for identifying leaf disease have been used in the past, research-based studies may produce a better result and overcome the disadvantages of the manual process. Nowadays, the random vector functional link (RVFL) network gaining popularity day by day due to the better generalization ability for various applications. In the spirit of the popular twin support vector machine (TSVM) and the twin variant of the random vector functional link (TRVFL) network, a novel twin RVFL (TRVFL1norm) is suggested in this work. TRVFL1norm overcomes the inherent limitation of TRVFL, which loses the sparsity due to the use of 2 norm and also gives a high penalty to outlier points which leads to degrade the generalization performance. Moreover, it is well known that Universum data can be used as prior knowledge in the training model. Therefore, to make the model more robust and generalized, we propose another model, TRVFL1norm based on the Universum data (UTRVFL1norm). UTRVFL1norm possesses the advantages of both 1-norm as well as Universum data. For both TRVFL1norm and UTRVFL1norm, few of the components in the output matrix will be zero due to sparsity hence, the resultant classifier must depend on the lesser output nodes compared to TRVFL. To verify model’s performance, we have compared the classification performance of proposed TRVFL1norm and UTRVFL1norm models with support vector machine (SVM), Universum SVM (USVM), twin support vector machine (TSVM), Universum TSVM (UTSVM), random vector functional link (RVFL), multilayer perceptron (MLP), convolutional neural network (CNN) and recurrent neural network (RNN) on few interesting benchmark datasets. Further, we have also checked the performance of TRVFL1norm and UTRVFL1norm for leaf disease detection as one of the possible applications. The experimental results indicate the efficiency and applicability of the proposed models.

Interplay of degeneracy and non-degeneracy in fluctuation propagation in coherent feed-forward loop motif

We present a stochastic framework to decipher fluctuation propagation in classes of coherent feed-forward loops (CFFLs). The systematic contribution of the direct (one-step) and indirect (two-step) pathways is considered to quantify fluctuations of the output node. We also consider both additive and multiplicative integration mechanisms of the two parallel pathways (one-step and two-step). Analytical expression of the output node’s coefficient of variation shows contributions of intrinsic, one-step, two-step, and cross-interaction in closed form. We observe a diverse range of degeneracy and non-degeneracy in each of the decomposed fluctuation terms and their contribution to the overall output fluctuations of each CFFL motif. The analysis of output fluctuations reveals a maximal level of fluctuations of the CFFL motif of type 1.

Food processing as the dominant contributor of phosphorus output in nearly half of China's regions

Human activities have rapidly increased the one-way flow in phosphorus cycle, leading to concerns about resource depletion and eutrophication, making it necessary to comprehensively analyze the phosphorus flows in the socio-economic system. This study used substance flow analysis to create a phosphorus accounting framework for China's natural and artificial systems, tracking 110 flows of 137 products in 31 provinces in 2020. The results show that mining is the largest phosphorus output node in China, followed by the phosphorus chemical industry and food processing industry. However, the dominant industry sectors in almost half of the regions are the food processing industry. The two phosphorus loss hotspots in China are the Yangtze River Basin dominated by human activities and the western regions dominated by natural processes. Considering phosphorus utilization, China does not face a significant disadvantage when compared to developed countries. These conclusions can assist in efficient phosphorus categorization and management.

Pretraining Graph Neural Networks for Few-Shot Analog Circuit Modeling and Design

Being able to predict the performance of circuits without running expensive simulations is a desired capability that can catalyze automated design. In this paper, we present a supervised pretraining approach to learn circuit representations that can be adapted to new circuit topologies or unseen prediction tasks. We hypothesize that if we train a neural network (NN) that can predict the output direct current (DC) voltages of a wide range of circuit instances it will be forced to learn generalizable knowledge about the role of each circuit element and how they interact with each other. The dataset for this supervised learning objective can be easily collected at scale since the required DC simulation to get ground truth labels is relatively cheap. This representation would then be helpful for few-shot generalization to unseen circuit metrics that require more time consuming simulations for obtaining the ground-truth labels. To cope with the variable topological structure of different circuits we describe each circuit as a graph and use graph neural networks (GNNs) to learn node embeddings. We show that pretraining GNNs on prediction of output node voltages can encourage learning representations that can be adapted to new unseen topologies or prediction of new circuit level properties with up to 10x more sample efficiency compared to a randomly initialized model. We further show that we can improve sample efficiency of prior SoTA model-based optimization methods by 2x (almost as good as using an oracle model) via fintuning pretrained GNNs as the feature extractor of the learned models.

A deep learning approach for polyline and building simplification based on graph autoencoder with flexible constraints

ABSTRACT Polyline and building simplification remain challenging in cartography. Most proposed algorithms are geometric-based and rely on specific rules. In this study, we propose a deep learning approach to simplify polylines and buildings based on a graph autoencoder (GAE). The model receives the coordinates of line vertices as inputs and obtains a simplified representation by reconstructing the original inputs with fewer vertices through pooling, in which the graph convolution based on graph Fourier transform is used for the layer-by-layer feature computation. By adjusting the loss functions, constraints such as area and shape preservation and angle-characteristic enhancement are flexibly configured under a unified learning framework. Our results confirmed the applicability of the GAE approach to the multi-scale simplification of land-cover boundaries and contours by adjusting the number of output nodes. Compared with existing Douglas‒Peukcer, Fourier transform, and Delaunay triangulation approaches, the GAE approach was superior in achieving morphological abstraction while producing reasonably low position, area, and shape changes. Furthermore, we applied it to simplify buildings and demonstrated the potential for preserving the diversified characteristics of different types of lines.

Incremental Multilayer Broad Learning System With Stochastic Configuration Algorithm for Regression

Broad learning system (BLS) is a novel randomized learning framework which has a faster modeling efficiency. Although BLS with incremental learning has a better extendibility for updating model rapidly, the incremental mode of BLS lacks self-supervision mechanism which cannot adjust the structure adaptively. Learning from the idea of stochastic configuration network (SCN), a novel incremental multi-layer broad learning system based on stochastic configuration (SC) algorithm is proposed for regression, termed as IMLBLS-SC. Firstly, to improve the feature learning ability, SC algorithm is adopted to configure the parameters of enhancement nodes instead of random weights. Secondly, the multi-layer model with enhancement nodes can be added gradually according to supervision mechanism without human intervention. Thirdly, all the enhancement nodes and feature nodes are fully connected with output nodes. Finally, 2 function approximation problems and 8 classical datasets are selected to verify the regression performance of IMLBLS-SC, experimental results demonstrate that IMLBLS-SC outperforms the RVFLN, SCN, BLS and BSCN.

Low Latency Energy-Efficient Neural Decoders for Block Codes

Neural decoders are being explored for 5G and beyond communication networks which have ultra-low-latency and energy-efficiency requirements. We consider several ultra-short linear and non-linear block codes and design neural decoders for them, which (i) provide near maximum-likelihood (ML) performance in terms of bit and block error rates, and (ii) have low decoding complexity (and hence, latency and energy consumption) in terms of the number of hidden layers and hidden nodes. We explore (i) single-label, (ii) multi-label, and (iii) concatenated coding based decoders. We find that single-label decoders with a single hidden layer having a sufficient number of nodes achieve near-ML performance. Since the number of output nodes required in single-label decoders is exponential in the message length, we investigate a multi-label neural decoder requiring a significantly lower number of output nodes but having decreased performance in terms of block error rate. To eliminate extra errors, we concatenate an outer code with the original code as the inner code. The inner and outer codes are decoded by multi-label and single-label neural decoders, respectively, leading to better block error performance than that is achievable using multi-label decoders.

A doubled transistor latch common-mode insensitive rail-to-rail regenerative comparator for low supply voltage applications

This paper presents a high-speed common-mode insensitive two-stage regenerative comparator designed for low supply voltage applications. The proposed comparator features a doubled transistor latch that enables rail-to-rail operation, which is pioneered to the best of the author's knowledge. The voltage swing of output nodes of the pre-amplifiers is limited to reduce the power consumption. This limitation boosts the regeneration speed of the doubled transistor latch. The proposed comparator realized by a standard 65 nm CMOS technology. The post-layout simulation results corroborate that the proposed comparator with 50 MHz sample rate and 0.6 V supply voltage dissipates 3.4 µW power, and occupies 168 µm2 area. Also, the offset voltage variation within the full common-mode range can be below 2 mV. In a worst-case scenario compared to the conventional single-stage comparator, the proposed comparator improves delay and energy efficiency by more than 80%. Considering its overall performance, this energy-efficient comparator is suitable for all low supply voltage applications, including Internet of Things (IoT), successive approximation registers analog-to-digital converters (SAR ADCs), and biomedical implants.

Subgraph Learning for Topological Geolocalization with Graph Neural Networks.

One of the challenges of spatial cognition, such as self-localization and navigation, is to develop an efficient learning approach capable of mimicking human ability. This paper proposes a novel approach for topological geolocalization on the map using motion trajectory and graph neural networks. Specifically, our learning method learns an embedding of the motion trajectory encoded as a path subgraph where the node and edge represent turning direction and relative distance information by training a graph neural network. We formulate the subgraph learning as a multi-class classification problem in which the output node IDs are interpreted as the object's location on the map. After training using three map datasets with small, medium, and large sizes, the node localization tests on simulated trajectories generated from the map show 93.61%, 95.33%, and 87.50% accuracy, respectively. We also demonstrate similar accuracy for our approach on actual trajectories generated by visual-inertial odometry. The key benefits of our approach are as follows: (1) we take advantage of the powerful graph-modeling ability of neural graph networks, (2) it only requires a map in the form of a 2D graph, and (3) it only requires an affordable sensor that generates relative motion trajectory.

Three-Stage Operational Amplifier with Frequency Compensation Using Cascade Zero

Short channel MOSFET exhibits the characteristics of wide bandwidth and low DC gain. A low DC gain causes a high gain error and a narrow output linear range in the closed loop. The DC gains can be improved by using the cascade structure, but frequency compensation is required due to the increase in the number of poles. The output nodes of each stage in a cascade Common-Source amplifier have a cascade of zero, and this zero is cancelled out by the input node of the next stage. This paper proposes a three-stage operational amplifier (op-amp) with frequency compensation using cascade zero. This op-amp was implemented in the 180 nm CMOS technology and achieved 86.96 MHz unity–gain frequency, 51.7° phase margin at 32 pF load capacitor and 99.83 dB DC gain, that is, a 36.21 dB improvement over a two-stage op-amp with the same power consumption. The op-amp consumed 7.74 mW with a supply voltage of 1.8 V.

Ultrasonic Spectacles for Blind Person

The ultrasonic spatial code line for blind people using Node MCU, web camera, ultrasonic sensor, and output of speaker is an innovative system designed to aid visually impaired individuals in navigating their environment independently and safely. The system works by using a combination of hardware components and software algorithms to provide real-time spatial feedback to the user. The system consists of a Node MCU microcontroller, which acts as the central processing unit and controls the flow of data between the various components. A web camera is used to capture a live video feed of the user's surroundings, which is then processed using image recognition algorithms to identify and locate obstacles in the environment. An ultrasonic sensor is used to measure the distance between the user and any obstacles detected in the video feed. This information is then used to generate a tactile feedback map that is transmitted to the user via a speaker output. The tactile feedback map consists of a series of vibrations that correspond to the location and distance of the obstacles in the user's environment. The ultrasonic spatial code line is a critical need for blind individuals who struggle with navigating their environment independently. The system provides a practical solution that can help improve their quality of life by enabling them to move around more safely and confidently. Additionally, the system is relatively low-cost and easy to implement, making it accessible to a broader range of individuals. Key Word: Electronic Travel Aid – Visual Impairment- Machine learning – Navigation Aid - Mobility – Ultrasonic spectacles – Blind navigation;

Target Threat Assessment in Air Combat with BP Neural Network for UAV

Target threat assessment is the most important element in command control communication and computer intelligence surveillance reconnaissance (C4ISR) system which is the key function in air combat situation analysis and can select the most valuable target to the commander. A threat assessment model of multi-UAV platform cooperative engagement built based on back propagate neural network (BPNN) is present in this paper, which provides a machine learning method for the threat assessment. The BPNN target threat assessment model has six input nodes, 10 hide nodes and 1 output node, which can fit the complicated nonlinear relationship between threat value and target behavior state. Simulation experiment is executed to verify the accuracy of the proposed BPNN target threat assessment method, and the result shows that the method is effective and can be applied to the real air combat.

Recognition of Single Japanese Sounds using Surface Electromyography Signals by Machine Learning

The method and results of recognizing single Japanese sounds using surface electromyography (sEMG) signals generated from muscles around the mouth are presented. We determined six features of the waveforms of four muscles (a total of 24 indexes) and recognized 45 single Japanese sounds. We used machine learning with a neural network to improve sound recognition. The neural network has a 24 node input layer, a 100 node intermediate layer, and a 45 node output layer. Each index of a sound was entered into the input layer; the probabilities of the sound were output to the output layer. They were compared, and the output (sound) with the highest probability was determined to be a recognition sound. We used the cross-entropy loss as the loss function and gradient descent as the machine learning method. Machine learning, which built a neural network, has dramatically increased recognition; it stands at approximately 94%.

A wide input range, external capacitor-less LDO with fast transient response

A high voltage, external capacitor-less low-dropout regulator (HVLDO) with transient enhancement loop is presented in this work. The proposed HVLDO is designed with high withstand voltage laterally-diffused MOS (LDMOS) transistors and a transient enhancement loop is proposed to properly inject/sink current to/from the gate and output nodes of the power transistors to achieve fast transient response and high stability. Fabricated in 0.5µm SOI BCD process, the HVLDO occupies an active area of 0.29mm2. Operating with an input voltage ranging from 5.2 to 20V, it supplies an output voltage of 5V and a maximum load of 100mA. For all load conditions, this design has the power supply rejection of -49dB@100kHz, the phase margin over 68.4deg and achieves a temperature coefficient of 13.15ppm/°C. Measurement results show that this design has a line regulation of 0.88mV/V and a load regulation of 0.22mV/mA. The proposed HVLDO features fast line transient response of 60/20mV@9.8V/µs, fast load transient response of 30/70mV@100mA/µs, and recovery time of 2µs without external capacitors. Compared with the prior art, this work achieves the best transient FOM of 12.19fs.

RF Multi-Functional Input-Reflectionless Dispersive-Delay Structure With Sharp-Rejection Filtering Using Channelization Techniques

A class of RF multi-functional input-reflectionless dispersive-delay structure (DDS) with linear-type in-band group-delay variation and sharp-rejection bandpass-filtering capability is reported. It exploits a two-branch-channelized/balanced-type circuit with similar low-order reflective DDS units inside its channels, which are connected through input/output 3-dB quadrature wideband couplers. The adopted DDS unit is based on a coupled-resonator network with a frequency-dependent cross-coupling. It introduces a pair of complex transmission zeros (TZs) to shape the intended in-band group-delay profile. Unequal transmission-line-based phase-shifting sections are also inserted at the outputs of both channels. In this manner, an input-absorptive behavior and selectivity enhancement with TZ generation are attained by means of transversal signal-interference techniques at the overall input and output nodes, respectively. Moreover, as the transfer-function phase term of the interference action in transmission is linear with frequency, the total group-delay pattern is defined by the DDS unit of the channels. As practical validation, a 1.5-GHz proof-of-concept microstrip prototype is constructed and tested.

Performance Enhancement of a Spin-Wave-Based Reservoir Computing System Utilizing Different Physical Conditions

We numerically study how to enhance reservoir computing performance by thoroughly extracting the spin-wave device potential for higher-dimensional information generation. The reservoir device has a 1-input exciter and 120-output detectors on the top of a continuous magnetic garnet film for spin-wave transmission. For various nonlinear and fading-memory dynamic phenomena distributing in the film space, small in-plane magnetic fields are used to prepare stripe domain structures and various damping constants at the film sides and bottom are explored. The ferromagnetic resonant frequency and relaxation time of spin precession clearly characterizes the change in spin dynamics with the magnetic field and damping constant. The common input signal for reservoir computing is a 1-GHz cosine wave with random 6-valued amplitude modulation. A basic 120-dimensional reservoir output vector is obtained from time-series signals at the 120-output detectors under each of three magnetic field conditions. Then, 240- and 360-dimensional reservoir output vectors are also constructed by concatenating two and three basic ones, respectively. In nonlinear autoregressive moving average (NARMA) prediction tasks, the computational performance is enhanced as the dimension of the reservoir output vector becomes higher and a significantly low prediction error is achieved for the tenth-order NARMA task using the 360-dimensional vector and optimum damping constant. The results are clear evidence that the collection of diverse output signals efficiently increases the dimensionality of the integrated reservoir state vector (i.e. reservoir-state richness) and thereby contributes to high computational performance. This paper demonstrates that performance enhancement through various configuration settings is a practical approach for on-chip reservoir computing devices with small numbers of real output nodes.

Study on ESD Protection Circuit by TCAD Simulation and TLP Experiment

The anti-ESD characteristic of the electronic system is paid more and more attention. Moreover, the on-chip electrostatic discharge (ESD) is necessary for integrated circuits to prevent ESD failures. In this paper, the mixed TCAD model of the ESD protection circuit is built and simulated, and the negative transmission line pulse (TLP) injection damage experiment is carried out on the CD4069UBC chip. The circuit model consists of three-dimensional shallow trench isolation (STI) diode TCAD models and a three-dimensional multi-gate Complementary Metal-Oxide-Semiconductor (CMOS) inverter TCAD model. Moreover, the TCAD modeling is based on a 0.25 μm technology node. Through the transient simulation of the electrothermal coupling, the electrical signal of the input and output nodes of the circuit and the distribution of the electrothermal parameters in the device are obtained. Moreover, by analyzing the simulation results, the physical phenomena and the mechanisms of interference and damage mechanism during TLP injection are explained. The location and type of diode damage in the TLP injection simulation results of the circuit model are consistent with the TLP experiment damage results. The proposed ESD protection circuit model and analysis method are beneficial to ESD robustness prediction and ESD soft damage analysis of IC.

Effective Reversal of Macrophage Polarization by Inhibitory Combinations Predicted by a Boolean Protein-Protein Interaction Model.

The function and polarization of macrophages has a significant impact on the outcome of many diseases. Targeting tumor-associated macrophages (TAMs) is among the greatest challenges to solve because of the low in vitro reproducibility of the heterogeneous tumor microenvironment (TME). To create a more comprehensive model and to understand the inner workings of the macrophage and its dependence on extracellular signals driving polarization, we propose an in silico approach. A Boolean control network was built based on systematic manual curation of the scientific literature to model the early response events of macrophages by connecting extracellular signals (input) with gene transcription (output). The network consists of 106 nodes, classified as 9 input, 75 inner and 22 output nodes, that are connected by 217 edges. The direction and polarity of edges were manually verified and only included in the model if the literature plainly supported these parameters. Single or combinatory inhibitions were simulated mimicking therapeutic interventions, and output patterns were analyzed to interpret changes in polarization and cell function. We show that inhibiting a single target is inadequate to modify an established polarization, and that in combination therapy, inhibiting numerous targets with individually small effects is frequently required. Our findings show the importance of JAK1, JAK3 and STAT6, and to a lesser extent STK4, Sp1 and Tyk2, in establishing an M1-like pro-inflammatory polarization, and NFAT5 in creating an anti-inflammatory M2-like phenotype. Here, we demonstrate a protein-protein interaction (PPI) network modeling the intracellular signalization driving macrophage polarization, offering the possibility of therapeutic repolarization and demonstrating evidence for multi-target methods.

Deterministic size discovery and topology recognition in radio networks with short labels

We consider size discovery and topology recognition in radio networks without collision detection, modeled by simple undirected connected graphs. The aim of size discovery is that all nodes output the number of nodes in the graph, and in the task of topology recognition, each node has to learn the topology of the graph and its position in it. Nodes are assigned labels which are (not necessarily different) binary strings. Each node uses its own label as input when executing the algorithm. The length of a labeling scheme is the largest length of a label. Our goal is to construct short labeling schemes for size discovery and topology recognition in arbitrary radio networks, and to design efficient deterministic distributed algorithms for each of these tasks, using these short schemes. In both cases, the length of our labeling schemes is asymptotically optimal.