Noise Tolerance Research Articles

Quick-connect scanning tunneling microscope head with nested piezoelectric coarse walkers.

To meet changing research demands, new scanning tunneling microscope (STM) features must constantly evolve. We describe the design, development, and performance of a modular plug-in STM, which is compact and stable. The STM head is equipped with a quick-connect socket that is matched to a universal connector plug, enabling it to be transferred between systems. This head can be introduced into a vacuum system via a load-lock and transferred to various sites equipped with the connector plug, permitting multi-site STM operation. Its design allows for reliable operation in a variety of experimental conditions, including a broad temperature range, ultra-high vacuum, high magnetic fields, and closed-cycle pulse-tube cooling. The STM's compact size is achieved by a novel nested piezoelectric coarse walker design, which allows for large orthogonal travel in the X, Y, and Z directions, ideal for studying both bulk and thin film samples ranging in size from mm to μm. Its stability and noise tolerance are demonstrated by achieving atomic resolution under ambient conditions on a laboratory desktop with no vibrational or acoustic isolation. The operation of the nested coarse walkers is demonstrated by successful navigation to a μm-sized 2D sample.

Inspection method of the corrosion rate for underwater grouting sleeves by integrating ultrasonic data augmentation and interpretable ensemble learning

To addresses the challenge of underwater grouting sleeve corrosion affecting bridge service life, we propose an innovative inspection approach by integrating a modified ultrasonic testing instrument, ultrasonic data augmentation method, and an improved ensemble learning prediction model. Firstly, the ultrasonic data was collected from grouting sleeve specimens by the modified instrument. Secondly, a data generation model, incorporating Bray Curtis distance and rejection sampling algorithm, was developed to generate samples and filter abnormal data. Then, Circle Chaotic Map improved Grey Wolf Optimizer was proposed to optimize parameters from the Random Forest (RF) algorithm for establishing the prediction model. Finally, SHapley Additive exPlanations (SHAP) method was used to interpret the model globally and locally. Experimental results validate the effectiveness of the data augmentation method, ensuring high-quality ultrasonic data for accurate corrosion rate predictions. The proposed inspection method provide technical support for assessing remaining bridge bearing capacity and service life, showcasing high accuracy, interpretability, and noise tolerance.

Predefined-time robust adaptive gradient neural network for solving linear time-varying equations and its applications

Adaptive gradient neural networks (AGNN) have been extensively employed as effective models for addressing time-varying problems. However, the existing AGNN are convergent in infinite time and only considered in noiseless environment. Consequently, the AGNN with a certain convergence time and noise tolerance is urgently required. Motivated by the advantages of predefined-time convergence, a novel predefined-time robust adaptive gradient neural network (PRAGNN) is developed for solving linear time-varying equations. It can be concluded that PRAGNN can efficiently obtain real-time solutions of linear time-varying equations within predefined time, independent of initial conditions, by properly selecting parameters. In the process of solving linear time-varying equations, PRAGNN can avoid matrix inversion, which reduces the computing cost compared with zeroing neural networks (ZNN). Furthermore, PRAGNN can completely resist both bounded vanishing noise and bounded non-vanishing noise. Numerical experiments demonstrate the effectiveness of PRAGNN. Finally, PRAGNN has been successfully utilized in 2-D dynamic positioning based on the angle of arrival (AOA) positioning algorithm, as well as in alternating currents (AC) computing.

Design of Tiny Contrastive Learning Network With Noise Tolerance for Unauthorized Device Identification in Internet of UAVs

Design of Tiny Contrastive Learning Network With Noise Tolerance for Unauthorized Device Identification in Internet of UAVs

INDIDO: A novel low-power approach for domino logic circuits

Power dissipation in Nanoelectronic circuits at advanced technology nodes is dominated by leakage power dissipation. This is due to an increase in short channel effects in the scaled transistors.The VLSI industry is seeking alternative options to enhance the performance of portable electronic systems by ensuring higher speed and reduced power dissipation. In the ultra-deep submicron (DSM) regime, a new device called the fin-shaped field effect transistor (FinFET) was developed to replace CMOS technology. FinFET, a multi-gate device, significantly reduces power dissipation compared to planer MOS (Metal Oxide Semiconductor ) transistor, but it doesn’t entirely resolve the issue. To further reduce power dissipation in the ultra DSM regime, low power approaches are required. Domino logic, a widely-used dynamic logic, is commonly utilized in high-speed VLSI architectures. In this research work, a novel INput Dependent Inverter DOmino (INDIDO) logic approach for low-power domino logic circuits using FinFET devices is proposed. A comparative analysis between the proposed INDIDO method and the existing approaches is performed for domino logic circuits for various performance metrics at the 16 nm technology node. In this research, various circuits like domino buffer, domino OR, domino AND, domino XOR and domino half adder are designed using the proposed INDIDO approach.The proposed INDIDO FinFET buffer circuit offers significant improvement in energy efficiency and FOM (Figure of Merit) by 53.18% and 82% respectively as compared to conventional FinFET footed domino buffer circuit. Beside this, proposed circuits like INDIDO OR, INDIDO AND, INDIDO XOR and INDIDO Half adder circuit follow the same trend as INDIDO buffer circuit and show better performance parametres in comparison to the existing low power domino approaches as well. In addition to this, the proposed INDIDO buffer circuits are also analyzed for PVT(process voltage and temperature) variability.This whole analysis makes us to conclude that proposed INDIDO approach reduces power dissipation and delay penalty, have high noise tolerance capacity, more immunity against PVT variations and high energy efficiency in comparison to the already existing techniques.

Noise tolerance of recycled entanglement detection by sequential and independent observers

Noise tolerance of recycled entanglement detection by sequential and independent observers

Is Acceptable Noise Level Affected by the Number of Competitive Talkers? An Implication of Informational Masking and Listening in Dips for Acceptable Noise Level Mechanism

Background and Aim: The Acceptable Noise Level (ANL), which is an effective clinical tool for quantitative assessment of noise tolerance, is affected by some known variables related to both subject and testing materials. The present study examined how the characteristics of different babble noises may affect the ANL results in normal adult listeners. Methods: Forty Persian listeners with normal hearing participated. In addition to typical ANL testing with 12-talker noise, the ANL was obtained in 8 different conditions varying in number of talkers from 2 to 10 in the babble noises presenting forward and backward. Results: There was a significantly lower ANL for 2-talker babble compared to 4, 8, 10, and 12-talker babble in both forward and backward noise conditions. With the increase in talkers in noise, the ANL becomes worse but reaches almost a plateau with more than 4 talkers in babble noise. There was a statistically significant difference between 2-talker forward and 2-talker backward noises, with no difference for the other conditions. Conclusion: This finding that the ANL is affected by the number of talkers in babble noise and by the forward and backward background noise suggests that informational masking and listening in dip mechanisms are involved in ANL for normal hearing people at least. Keywords: Acceptable noise level; speech babble noise; listening in dips; informational masking; energetic masking

Map retrieval intention recognition based on relevance feedback and geographic semantic guidance: For better understanding user retrieval demands

Effective retrieval is essential for finding resources in demand handily amidst extensive data records in data warehouse. Mainstream map retrieval methods suffer from intention gap problem and are incapable to describe sophisticated user demands precisely due to the limits of low- and middle-level text or visual feature matching, resulting in unsatisfactory retrieval results. Such limitations are more marked when map retrieval demands were characterized with joint constraints of geographic concepts. To address this issue, we propose a map retrieval intention recognition method to perceive user demands with relevance feedback samples and geographic semantics guidance. Specifically, we construct a hierarchical intention expression model to describe retrieval goals and their multi-dimensional attribute constrains; incorporate geographic ontologies to provide semantic guidance and facilitate recognition; utilize the frequent itemset mining (FIM) algorithm Apriori to generate intention candidates from relevance feedback samples, and search for the optimal intention set by adopting the minimum description length (MDL) principle. The experiments verify the effectiveness of Apriori algorithm and MDL principle on intention recognition. The proposed method outperforms the FIM algorithm Gene Ontology (RuleGO) and the Decision Tree algorithm with Hierarchical Features (DTHF) with higher recognition accuracy and noise tolerance. Furthermore, through our sample augmentation strategy, the method yields promising recognition accuracy even when the feedback sample size is as low as ten, substantially reducing the feedback burden in human-computer interactions. We envision that the application of our method in spatial data infrastructures (SDIs), such as geoportals and catalogue services, could enhance the quality of service and user experience in geospatial data discovery.

Robust deep image clustering using convolutional autoencoder with separable discrete Krawtchouk and Hahn orthogonal moments

By cooperatively learning features and assigning clusters, deep clustering is superior to conventional clustering algorithms. Numerous deep clustering algorithms have been developed for a variety of application levels; however, the majority are still incapable of learning robust noise-resistant latent features, which limits the clustering performance. To address this open research challenge, we introduce, for the first time, a new approach called: Robust Deep Embedded Image Clustering algorithm with Separable Krawtchouk and Hahn Moments (RDEICSKHM). Our approach leverages the advantages of Krawtchouk and Hahn moments, such as local feature extraction, discrete orthogonality, and noise tolerance, to obtain a meaningful and robust image representation. Moreover, we employ LayerNormalization to further improve the latent space quality and facilitate the clustering process. We evaluate our approach on four image datasets: MNIST, MNIST-test, USPS, and Fashion-MNIST. We compare our method with several deep clustering methods based on two metrics: clustering accuracy (ACC) and normalized mutual information (NMI). The experimental results show that our method achieves superior or competitive performance on all datasets, demonstrating its effectiveness and robustness for deep image clustering.

Measurement-free mediated semi-quantum key distribution protocol based on single-particle states

A novel measurement-free mediated semi-quantum key distribution (MSQKD) protocol is proposed based on single-particle states. It enables two classical users to establish a secret key with the assistance of a third party. This protocol simplifies the third party’s role to solely generating qubits in X-basis and conducting Bell measurements. A distinctive feature of this protocol is the efficient grouping and reordering of qubits by the classical users with a minimum of three delay lines. Security analyses demonstrate that the protocol can withstand various attack strategies, including collective attack, measurement attack, fake state attack, and modification attack. The noise tolerance is given by deriving a lower bound of the protocol’s key rate in the asymptotic scenario. Simulations on the IBM Quantum Experience platform are conducted to illustrate the feasibility of this protocol. Compared with existing MSQKD protocols, the proposed protocol consumes fewer quantum resources and achieves a qubit efficiency of 1/8.

Video Moment Retrieval With Noisy Labels.

Video moment retrieval (VMR) aims to localize the target moment in an untrimmed video according to the given nature language query. The existing algorithms typically rely on clean annotations to train their models. However, making annotations by human labors may introduce much noise. Thus, the video moment retrieval models will not be well trained in practice. In this article, we present a simple yet effective video moment retrieval framework via bottom-up schema, which is in end-to-end manners and robust to noisy label training. Specifically, we extract the multimodal features by syntactic graph convolutional networks and multihead attention layers, which are fused by the cross gates and the bilinear approach. Then, the feature pyramid networks are constructed to encode plentiful scene relationships and capture high semantics. Furthermore, to mitigate the effects of noisy annotations, we devise the multilevel losses characterized by two levels: a frame-level loss that improves noise tolerance and an instance-level loss that reduces adverse effects of negative instances. For the frame level, we adopt the Gaussian smoothing to regard noisy labels as soft labels through the partial fitting. For the instance level, we exploit a pair of structurally identical models to let them teach each other during iterations. This leads to our proposed robust video moment retrieval model, which experimentally and significantly outperforms the state-of-the-art approaches on standard public datasets ActivityCaption and textually annotated cooking scene (TACoS). We also evaluate the proposed approach on the different manual annotation noises to further demonstrate the effectiveness of our model.

Noisecut: a python package for noise-tolerant classification of binary data using prior knowledge integration and max-cut solutions

BackgroundClassification of binary data arises naturally in many clinical applications, such as patient risk stratification through ICD codes. One of the key practical challenges in data classification using machine learning is to avoid overfitting. Overfitting in supervised learning primarily occurs when a model learns random variations from noisy labels in training data rather than the underlying patterns. While traditional methods such as regularization and early stopping have demonstrated effectiveness in interpolation tasks, addressing overfitting in the classification of binary data, in which predictions always amount to extrapolation, demands extrapolation-enhanced strategies. One such approach is hybrid mechanistic/data-driven modeling, which integrates prior knowledge on input features into the learning process, enhancing the model’s ability to extrapolate.ResultsWe present NoiseCut, a Python package for noise-tolerant classification of binary data by employing a hybrid modeling approach that leverages solutions of defined max-cut problems. In a comparative analysis conducted on synthetically generated binary datasets, NoiseCut exhibits better overfitting prevention compared to the early stopping technique employed by different supervised machine learning algorithms. The noise tolerance of NoiseCut stems from a dropout strategy that leverages prior knowledge of input features and is further enhanced by the integration of max-cut problems into the learning process.ConclusionsNoiseCut is a Python package for the implementation of hybrid modeling for the classification of binary data. It facilitates the integration of mechanistic knowledge on the input features into learning from data in a structured manner and proves to be a valuable classification tool when the available training data is noisy and/or limited in size. This advantage is especially prominent in medical and biomedical applications where data scarcity and noise are common challenges. The codebase, illustrations, and documentation for NoiseCut are accessible for download at https://pypi.org/project/noisecut/. The implementation detailed in this paper corresponds to the version 0.2.1 release of the software.

Sensitivity volume as figure-of-merit for maximizing data importance in electrical impedance tomography

Objective. Electrical impedance tomography (EIT) is a noninvasive imaging method whereby electrical measurements on the periphery of a heterogeneous conductor are inverted to map its internal conductivity. The EIT method proposed here aims to improve computational speed and noise tolerance by introducing sensitivity volume as a figure-of-merit for comparing EIT measurement protocols. Approach. Each measurement is shown to correspond to a sensitivity vector in model space, such that the set of measurements, in turn, corresponds to a set of vectors that subtend a sensitivity volume in model space. A maximal sensitivity volume identifies the measurement protocol with the greatest sensitivity and greatest mutual orthogonality. A distinguishability criterion is generalized to quantify the increased noise tolerance of high sensitivity measurements. Main result. The sensitivity volume method allows the model space dimension to be minimized to match that of the data space, and the data importance to be increased within an expanded space of measurements defined by an increased number of contacts. Significance. The reduction in model space dimension is shown to increase computational efficiency, accelerating tomographic inversion by several orders of magnitude, while the enhanced sensitivity tolerates higher noise levels up to several orders of magnitude larger than standard methods.

Characterization of the ATLAS Liquid Argon Front-End ASIC ALFE2 for the HL-LHC upgrade

ALFE2 is an ATLAS Liquid Argon Calorimeter (LAr) Front-End ASIC designed for the HL-LHC upgrade. ALFE2 comprises four channels of pre-amplifiers and CR-(RC)2 shapers with adjustable input impedance. ALFE2 features two separate gain outputs to provide 16-bit dynamic-range coverage and an optimum resolution. ALFE2 is characterized using a Front-End Test Board (FETB) based on a Zynq UltraScale+ MPSoC and two octal-channel 16-bit high-speed ADCs. The test results indicate that ALFE2 fulfills or greatly exceeds all specifications on gain, noise, linearity, uniformity, and radiation tolerance.

A Convergence Path to Deep Learning on Noisy Labels.

In many real-world machine learning classification applications, the model performance based on deep neural networks (DNNs) oftentimes suffers from label noise. Various methods have been proposed in the literature to address this issue, primarily by focusing on designing noise-tolerant loss functions, cleaning label noise, and correcting the objective loss. However, the noise-tolerant loss functions face challenges when the noise level increases. This article aims to reveal a convergence path of a trained model in the presence of label noise, and here, the convergence path depicts the evolution of a trained model over epochs. We first propose a theorem to demonstrate that any surrogate loss function can be used to learn DNNs from noisy labels. Next, theories on the general convergence path for the deep models under label noise are presented and verified through a series of experiments. In addition, we design an algorithm based on the proposed theorems that make efficient corrections on the noisy labels and achieve strong robustness in the DNN models. We designed several experiments using benchmark datasets to assess noise tolerance and verify the theorems presented in this article. The comprehensive experimental results firmly confirm our theoretical results and also clearly validate the effectiveness of our method under various levels of label noise.

Flame soot absorption coefficient and temperature reconstruction through line-of-sight attenuation and light field imaging

The accuracy and efficiency of 3-D flame temperature field reconstruction using the light field imaging technique heavily depend on soot radiation characteristics. In this study, we employ the line-of-sight attenuation method to reconstruct the soot absorption coefficient distribution in a pure absorbing flame. Utilizing these soot absorption coefficients, the radiative transfer equation is effectively transformed within the framework of the light field imaging technique into a linear inverse problem and also outlines the flame boundary. This proposed strategy reduces the unwanted detection rays significantly, thus eliminating the extensive computational processing. Consequently, the proposed approach substantially enhances the accuracy and efficiency of flame temperature reconstruction. Numerical simulations were carried out on a bimodal asymmetric flame to validate the noise tolerance capabilities, investigate the effects of varying voxel numbers on flame division and carry out a comparative study. Experimental studies were also conducted to reconstruct flame temperature and soot absorption coefficient distributions under different combustion operating conditions. Thermocouple measurements were performed and compared with the reconstructed temperatures. Outcomes achieved from both numerical simulations and experimental studies demonstrate the feasibility, accuracy and robustness of the proposed method.

An Integrated All‐Optical Ising Machine with Unlimited Spin Array Size and Coupling

Artificial spin systems are used to solve combinatorial optimization problems by mapping them to the ground state search of the Ising model. Ising machines of various designs, in fields as diverse as photonics and electronics, have been proposed and demonstrated in recent years. One important mathematical operation required in photonic Ising machines for nearly all Hamiltonian minimization algorithms is repeated matrix–vector multiplication with the vector size limited by the physical size of the optical system. Herein, an integrated photonic Ising machine based on matrix partitioning and phase encoding, which effectively allows the use of physically small optical circuits to handle arbitrarily large spin vectors, is proposed. All the complicated calculations are carried out optically, rather than electronically, in this approach. How the required surface area of a hypothetical chip‐based photonic Ising machine can be resized according to convenience, with the trade‐off being the solution time, and that this trade‐off can be done in a way that does not impact the algorithm's efficacy, is shown. Through simulation, the system is predicted to have a high success rate, high noise tolerance, and high error tolerance.

Fully visible light controlled neuro-synaptic ReRAM device based on metal oxide heterojunction

Photoelectric synaptic devices as a combination of electronic synapse and photodetector are considered as emerging bio-inspired device technologies. These devices have immense potential to conquer the bottleneck of von Neumann architecture based traditional computing systems. In this Letter, we propose an all-oxide based photoelectric neuro-synaptic resistive random access memory device with the structure of ITO/Ga2O3/ZnO/ITO/Glass, in which the conductance states are reversibly tuned by two different wavelengths (405/522 nm) of visible light spectrum. The strength of light pulse is altered to investigate the learning and forgetting phases of the photoelectric response of the device. A basic biomimetic function “learning-forgetting-rehearsal” behavior is imitated up to 20 cycles. Moreover, emulation of some typical synaptic functions such as associative learning and switching between short and long term plasticities indicate the wavelength awareness of the device. Based on the pure optically induced potentiation/depression characteristics, convolutional neural network simulation achieves an overall test accuracy of 82.5% for the classification of Zalando's article images. The noise tolerance capability of neural network is also examined by applying “salt and pepper” noise in high proportion (75%) to corrupt the images. This work may provide a promising step toward the development of transparent electronics in optogenetics-inspired neuromorphic computing.

Phase noise matching in resonant metasurfaces forintrinsic sensing stability.

Interferometry offers a precise means of interrogating resonances in dielectric and plasmonic metasurfaces, surpassing spectrometer-imposed resolution limits. However, interferometry implementations often face complexity or instability issues due to heightened sensitivity. Here, we address the necessity for noise compensation and tolerance by harnessing the inherent capabilities of photonic resonances. Our proposed solution, termed "resonant phase noise matching," employs optical referencing to align the phases of equally sensitive, orthogonal components of the same mode. This effectively mitigates drift and noise, facilitating the detection of subtle phase changes induced by a target analyte through spatially selective surface functionalization. Validation of this strategy using Fano resonances in a 2D photonic crystal slab showcases noteworthy phase stability (). With demonstrated label-free detection of low-molecular-weight proteins at clinically relevant concentrations, resonant phase noise matching presents itself as a potentially valuable strategy for advancing scalable, high-performance sensing technology beyond traditional laboratory settings.

Relationship between intra-community noise tolerance distributions and fitted dose-response funtions

Increasingly, researchers are employing fitted exposure-response functions to not only characterize the relationship between sound level and prevalence of annoyance but also to compare the relative noise tolerance of individual communities. The sound level at which 50% of the population is annoyed has been advanced as a community noise tolerance metric. This paper combines prior research and recent simulations to reframe the discussion in terms of plausible distributions of individual (personal) noise tolerances within a given community; the integrals of those (density) distributions become the (cumulative) observed exposure-response relationships. Various tolerance distributions are developed, their resulting cumulative exposure-response relationships calculated, and the ability of commonly-employed functional forms to represent those cumulative relationships presented. The results of the current investigation strongly suggest that data sets which include annoyance fractions of 50% have the greatest chance of accurately predicting the 50%-annoyed sound level. The further the data set lies from the 50%-annoyed point the more dependent an extrapolated estimation becomes on the fitted functional form accurately representing the true cumulative relationship. Errors of estimation are shown for a number of tolerance distributions, functional forms and experimental data ranges. Finally, some remedial recommendations are offered.