Accuracy Degradation Research Articles

Discriminative embedded multi-view fuzzy C-means clustering for feature-redundant and incomplete data

Multi-view clustering is a widely-used technique that seeks to categorize data obtained from various sources. As a representative method, multi-view fuzzy clustering has attracted growing attention. However, it becomes quite challenging when feature-redundant and incomplete data is presented. Despite the existing studies on dimension reduction and imputation methods, several issues remain unresolved. There is an excessive concern on the imputation, without considering that interpolation methods lead to accuracy degradation. Moreover, most of the methods usually process these two steps separately, resulting in inefficiency. To address these issues, we propose a discriminative embedded incomplete multi-view fuzzy c-means clustering method. We construct the indicator matrix to guide the learning of the common membership function, and design the projection matrix to construct embedding spaces. Subsequently, we develop an iterative optimization algorithm that solves the resultant problem. We demonstrate that the projection matrix can be achieved through the utilization of eigenvalue decomposition. Through extensive experimental studies on various benchmark datasets, the proposed method demonstrates the effectiveness and efficiency compared to the existing state-of-the-art clustering algorithms.

Cross-condition bearing fault detection based on online drift detection and domain adaptation

Aiming at the problem that the data distribution of bearings across operating conditions generates offset resulting in insufficient diagnostic accuracy of the original model for new data, a cross-condition bearing fault detection method based on online drift detection and domain adaptation is proposed. First, the original one-dimensional vibration signals collected are transformed by a two-dimensional wavelet transform to convert the time-frequency image dataset. Second, the drift detection of the data across operating conditions is carried out using Random Forest (RF), and the 3σ criterion as well as the drift detection judgment criteria are set. Next, the source domain model based on Googlenet is used to extract features from the target domain data, and the Whale Optimization Algorithm to Improve Local Preserving Projection Algorithm (WOA-LPP) algorithm is combined to construct a brand-new projection space to align the features of the source and target domains. Then, the source and target domain features are reconstructed by combining the LPP optimal projection matrix to construct a fully connected network trained by the source domain features. Finally, probabilistic label-based decision fusion is proposed to integrate multiple classifiers to reduce the effects of model training randomness and strong noise interference. Validated by the publicly available Western Reserve University bearing data, the method proposed in this paper has good detection accuracy as well as robustness across operating conditions, which can effectively improve the defects of shifting data distribution and degradation of model accuracy under variable speed.

Correlation between rate of TEC index and positioning error during solar flares and geomagnetic storms using navigation with Indian constellation receiver measurements

Abstract The real-time position accuracy of the Global Navigation Satellite System (GNSS)/Navigation with Indian Constellation (NavIC) receiver is limited by the dynamic behavior of the ionosphere, particularly in adverse conditions like solar flares and geomagnetic storms. The NavIC satellites broadcast dual coherent radio beacon signals on L5 (1,164.5 MHz) and S (2,472.5 MHz) bands for providing position, velocity, and timing services in all weather conditions. The Total Electron Content (TEC) and Rate of TEC Index (ROTI) are the potential indicators for characterizing the ionosphere and its irregularities. In this research work, the TEC and ROTI are computed from the code and carrier phase observations of the NavIC receiver located at Kurnool low latitude station (15.79° N, 78.07° E) with geomagnetic coordinates (7.30° N, 151.65° E). This paper presents a statistical study of TEC, ROTI, and the correlation between ROTI and NavIC positioning error during highly intense solar flares (X9.3 and X2.2) and geomagnetic storm conditions. Compared to quiet days mean TEC, the enhancement is 3 TECU due to X9.3 flares, and the maximum peak of TEC on storm day (September 8, 2017) is 80.92 TECU. Moreover, the correlation coefficient between ROTI and position error is 0.76 on a quiet day (September 4, 2017), 0.54 on an intense solar flares day (September 6, 2017), and 0.24 on a storm day (September 8, 2017), this indicates positional accuracy degradation on a geomagnetic storm day. The outcome of this research work would be helpful for investigating characteristics of the northern low latitude ionospheric irregularities and, in turn, useful for developing suitable ionospheric nowcasting/prediction models for GNSS applications.

Physics-Informed Neural Operator for Learning Partial Differential Equations

In this article, we propose physics-informed neural operators (PINO) that combine training data and physics constraints to learn the solution operator of a given family of parametric Partial Differential Equations (PDE). PINO is the first hybrid approach incorporating data and PDE constraints at different resolutions to learn the operator. Specifically, in PINO, we combine coarse-resolution training data with PDE constraints imposed at a higher resolution. The resulting PINO model can accurately approximate the ground-truth solution operator for many popular PDE families and shows no degradation in accuracy even under zero-shot super-resolution, that is, being able to predict beyond the resolution of training data. PINO uses the Fourier neural operator (FNO) framework that is guaranteed to be a universal approximator for any continuous operator and discretization convergent in the limit of mesh refinement. By adding PDE constraints to FNO at a higher resolution, we obtain a high-fidelity reconstruction of the ground-truth operator. Moreover, PINO succeeds in settings where no training data is available and only PDE constraints are imposed, while previous approaches, such as the Physics-Informed Neural Network (PINN), fail due to optimization challenges, for example, in multi-scale dynamic systems such as Kolmogorov flows.

Deep network fault diagnosis for imbalanced small-sized samples via a coupled adversarial autoencoder based on the Bayesian method.

Deep network fault diagnosis methods heavily rely on abundant labeled data for effective model training. However, small-sized samples and imbalanced samples often lead to insufficient features, resulting in accuracy degradation and even instability in the diagnosis model. To address this challenge, this paper introduces a coupled adversarial autoencoder (CoAAE) based on the Bayesian method. This model aims to solve the issue of insufficient samples by generating fake samples and integrating them with the original ones. Within the CoAAE framework, the probability density distribution of the original data is captured using an encoder and fake samples are generated by random sampling from this distribution and decoding them. This process is the adversarial interaction between the encoder and a classifier to obtain the prior distribution of the encoder's parameters. The encoder's parameters are updated through the decoder's reconstruction process, leading to the posterior distribution. Concurrently, the decoder is trained to enhance its ability to reconstruct samples accurately. To address the imbalance in the original samples, a parallel coupled network is employed. This network shares the weights of the extraction layer in the encoder, enabling it to learn the joint distribution between fault-related and normal samples. To evaluate the effectiveness of the proposed data augmentation method, experiments were conducted on a bearing database from Case Western Reserve University using ResNet18 as the deep learning diagnosis model representative. The results demonstrate that CoAAE can effectively augment imbalanced datasets and outperform other advanced methods.

DLformer: A Dynamic Length Transformer-Based Network for Efficient Feature Representation in Remaining Useful Life Prediction.

Representation learning-based remaining useful life (RUL) prediction plays a crucial role in improving the security and reducing the maintenance cost of complex systems. Despite the superior performance, the high computational cost of deep networks hinders deploying the models on low-compute platforms. A significant reason for the high cost is the computation of representing long sequences. In contrast to most RUL prediction methods that learn features of the same sequence length, we consider that each time series has its characteristics and the sequence length should be adjusted adaptively. Our motivation is that an "easy" sample with representative characteristics can be correctly predicted even when short feature representation is provided, while "hard" samples need complete feature representation. Therefore, we focus on sequence length and propose a dynamic length transformer (DLformer) that can adaptively learn sequence representation of different lengths. Then, a feature reuse mechanism is developed to utilize previously learned features to reduce redundant computation. Finally, in order to achieve dynamic feature representation, a particular confidence strategy is designed to calculate the confidence level for the prediction results. Regarding interpretability, the dynamic architecture can help human understand which part of the model is activated. Experiments on multiple datasets show that DLformer can increase up to 90% inference speed, with less than 5% degradation in model accuracy.

Refining computer tomography data with super-resolution networks to increase the accuracy of respiratory flow simulations

Accurately computing the flow in the nasal cavity with computational fluid dynamics (CFD) simulations requires highly resolved computational meshes based on anatomically realistic geometries. Such geometries can only be obtained from computer tomography (CT) data with high spatial resolution, i.e., featuring a ≤1mm slice thickness. In practice, CT images are, however, recorded at a lower resolution to not expose patients to high radiation and to reduce the overall costs. To overcome this problem and to provide patients with a detailed physics-based diagnosis, e.g., for surgery planning, the potential of super-resolution networks (SRNs) to increase the CT resolution is analyzed. Therefore, an SRN is developed and trained on CT data. Its predictive performance is improved by an automated hyperparameter optimization technique. The training time is further reduced without predictive accuracy degradation by oversampling images with challenging regions. The performance of the SRN is assessed by an analysis of the reconstructed 3D surfaces of the human upper airway and by comparing results of CFD simulations. That is, surfaces and simulation results based on SRN-generated CT data at 1mm resolution are compared to those obtained from unmodified CT data-sets at low (3mm) and high (1mm) resolution, as well as from CT data interpolated to a 1mm resolution from coarse data. The findings reveal the SRN-based approach to have the lowest deviations in the surfaces and CFD results when compared to those based on the original high-resolution data. The pressure loss between the inflow (nostrils) and outflow (pharynx) regions averaged for three test patients differs by only 1.3%, compared to 8.7% and 8.8% in the coarse and interpolated cases. It is concluded that the SRN-based method is a promising tool to enhance underresolved CT data to yield reliable numerical results of respiratory flows.

Toward Practical Superconducting Accelerators for Machine Learning Using U-SFQ

Most popular superconducting circuits operate on information carried by ps-wide, μV-tall, single flux quantum (SFQ) pulses. These circuits can operate at frequencies of hundreds of GHz with orders of magnitude lower switching energy than complementary-metal-oxide-semiconductors (CMOS). However, under the stringent area constraints of modern superconductor technologies, fully-fledged, CMOS-inspired superconducting architectures cannot be fabricated at large scales. Unary SFQ (U-SFQ) is an alternative computing paradigm that can address these area constraints. In U-SFQ, information is mapped to a combination of streams of SFQ pulses and in the temporal domain. In this work, we extend U-SFQ to introduce novel building blocks such as a multiplier and an accumulator. These blocks reduce area and power consumption by 2\(\times\)and 4\(\times\)compared with previously proposed U-SFQ building blocks and yield at least 97% area savings compared with binary approaches. Using these multiplier and adder, we propose a U-SFQ Convolutional Neural Network (CNN) hardware accelerator capable of comparable peak performance with state-of-the-art superconducting binary approach (B-SFQ) in 32\(\times\)less area. CNNs can operate with 5–8 bits of resolution with no significant degradation in classification accuracy. For 5 bits of resolution, our proposed accelerator yields 5\(\times\)to 63\(\times\)better performance than CMOS and 15\(\times\)to 173\(\times\)better area efficiency than B-SFQ.

Characterization of Sand and Dust Pollution Degradation Based on Sensitive Structure of Microelectromechanical System Flow Sensor.

The effect of sand and dust pollution on the sensitive structures of flow sensors in microelectromechanical systems (MEMS) is a hot issue in current MEMS reliability research. However, previous studies on sand and dust contamination have only searched for sensor accuracy degradation due to heat conduction in sand and dust cover and have yet to search for other failure-inducing factors. This paper aims to discover the other inducing factors for the accuracy failure of MEMS flow sensors under sand and dust pollution by using a combined model simulation and sample test method. The accuracy of a flow sensor is mainly reflected by the size of its thermistor, so in this study, the output value of the thermistor value was chosen as an electrical characterization parameter to verify the change in the sensor's accuracy side by side. The results show that after excluding the influence of heat conduction, when sand particles fall on the device, the mutual friction between the sand particles will produce an electrostatic current; through the principle of electrostatic dissipation into the thermistor, the principle of measurement leads to the resistance value becoming smaller, and when the sand dust is stationary for some time, the resistance value returns to the expected level. This finding provides theoretical guidance for finding failure-inducing factors in MEMS failure modes.

Reliability of probes for CNC machine tools

So far, the issue of degradation of touch-trigger probe accuracy during measurements on CNC machines has not been addressed. In the article, for the first time, a procedure for testing and predicting the accuracy and reliability of probes used in machine tools was proposed and experimentally verified. Reliability test procedure covering 30 million triggers was performed under laboratory conditions. Based on the obtained results, a concept of the probe reliability testing procedure was proposed, and an attempt was made to assess the impact of the operational process on the probe's metrological parameters. An important issue of this work is to develop a concept for the analysis and interpretation of the results of such a process by determining which parameters of the probe description provide the most information about the degradation process and how to predict the operating time of such a device.

Compensation architecture design utilizing residual resource to mitigate impacts of nonidealities in RRAM-based computing-in-memory chips

Resistive random access memory (RRAM) is a promising technology for energy-efficient in-memory computing. However, due to technology limits, RRAM device faces a series of reliability issues. Deep neural network (DNN) computing based on RRAM suffers from accuracy degradation. On the one hand, offline DNN training solutions are difficult to fully consider and simulate all nonidealities. Worse still, new error or nonideality may come up with the usage of RRAM, which further deteriorates the effectiveness of offline training. On the other hand, online training poses great challenges on programming overhead and device lifetime. The iterative write-verify technique to program multi-bit RRAM cells prolongs write latency more than 10× longer than read latency. To overcome these issues, we propose a compensation architecture and training designs to mitigate the realistic accuracy loss in RRAM chips with negligible hardware resource for RRAM-based DNN computing. Firstly, we add trainable compensation channels in crossbars utilizing the residual resource after original weight mapping. Secondly, an offline training procedure with computing output from hardware is triggered to settle down appropriate weight value in compensation channels. Experimental results demonstrate that the proposed design can guarantee ≤ 0.8% loss of accuracy in DNN even when nonidealities reduce the original accuracy down to ≤73%.

Volume Structure Retrieval Using Drone-Based SAR Interferometry with Wide Fractional Bandwidth

Synthetic aperture radar (SAR) interferometry (InSAR) is a well-established remote sensing technique capable of providing accurate topographic information of an area. Multiple scattering occurring at different heights within a single resolution cell, however, cannot be resolved using a single baseline and results in a degradation of the height accuracy. Techniques such as polarimetric InSAR and SAR tomography tackle this problem using additional measurements to obtain the three-dimensional structure of the volume. However, polarimetry-based methods assume orthogonal and deterministic scattering mechanisms and tomography requires numerous observations. This paper presents a novel approach to retrieving the three-dimensional structure of semitransparent media using wide fractional bandwidth InSAR. The frequency dependency of the InSAR coherence is exploited to retrieve the three-dimensional structure, assuming that signals with wide fractional bandwidth are available, as is the case for unmanned aerial vehicles (UAVs) or drones. As the extinction is, in general, frequency-dependent, a third observation may improve the inversion results. Simulations of different scenarios show that the frequency profile can be inverted to obtain information about the three-dimensional scattering structure. The acquired data also enable retrieval of the phase center depth as a function of frequency. This method allows for frequent and accurate monitoring of semitransparent targets, e.g., forests and vegetation, using a swarm of two or three drones.

A new compact scheme-based Lax–Wendroff method for high fidelity simulations

The classical Lax–Wendroff (LW) method employs explicit second order central difference schemes to solve partial differential equations. Recently, the authors in Sengupta et al. (2023) have identified numerical parameter ranges in performing large eddy simulation (LES) using this classical LW method. In a bid to improve these numerical parameter ranges for high fidelity computations, a new non-uniform grid based compact scheme for LW method (NUCLW) is developed here. This new method uses sixth order non-uniform compact scheme (NUC6) for the first and higher order derivatives. Reported global spectral analysis confirms substantial improvement in the resolution and accuracy of NUCLW scheme over the classical LW scheme. A significant advantage of the present scheme is its ability to compute solutions for non-uniform, structured grids, and that does not suffer degradation in accuracy. The potential of the NUCLW scheme for high accuracy computations is demonstrated by solving 2D incompressible Navier–Stokes equations for the benchmark unsteady flow inside a square lid driven cavity problem for a subcritical Reynolds number of 3200 and a post-critical Reynolds number of 10,000 and simulating the instability of the Taylor–Green vortex problem. These demonstrate the ability of the NUCLW scheme to solve the Navier–Stokes equations for high fidelity simulations such as LES/DNS with enhanced ranges of numerical parameters with respect to enhanced accuracy and faster computations.

Patient-independent, MHD-robust R-peak detection for retrospective gating in cardiac MRI imaging

Objective. In cardiovascular magnetic resonance imaging, synchronization of image acquisition with heart motion (called gating) is performed by detecting R-peaks in electrocardiogram (ECG) signals. Effective gating is challenging with 3T and 7T scanners, due to severe distortion of ECG signals caused by magnetohydrodynamic effects associated with intense magnetic fields. This work proposes an efficient retrospective gating strategy that requires no prior training outside the scanner and investigates the optimal number of leads in the ECG acquisition set. Approach. The proposed method was developed on a data set of 12-lead ECG signals acquired within 3T and 7T scanners. Independent component analysis is employed to effectively separate components related with cardiac activity from those associated to noise. Subsequently, an automatic selection process identifies the components best suited for accurate R-peak detection, based on heart rate estimation metrics and frequency content quality indexes. Main results. The proposed method is robust to different B0 field strengths, as evidenced by R-peak detection errors of 2.4 ± 3.1 ms and 10.6 ± 15.4 ms for data acquired with 3T and 7T scanners, respectively. Its effectiveness was verified with various subject orientations, showcasing applicability in diverse clinical scenarios. The work reveals that ECG leads can be limited in number to three, or at most five for 7T field strengths, without significant degradation in R-peak detection accuracy. Significance. The approach requires no preliminary ECG acquisition for R-peak detector training, reducing overall examination time. The gating process is designed to be adaptable, completely blind and independent of patient characteristics, allowing wide and rapid deployment in clinical practice. The potential to employ a significantly limited set of leads enhances patient comfort.

A Privacy-Aware and Incremental Defense Method Against GAN-Based Poisoning Attack

Federated learning is usually utilized as a fraud detection framework in the domain of financial risk management, which promotes the model accuracy without training data exchange. One of the challenges in federated learning is the GAN-based poisoning attack. The GAN-based poisoning attack is a type of intractable poisoning attack that causes global model accuracy degradation and privacy leak. Most of the existing defenses for GAN-based poisoning attack have the three problems: 1) dependence on validation datasets; 2) incompetence of dealing with incremental poisoning attack; and 3) privacy leak. To address the above problems, we present a privacy-aware and incremental defense (PID) method to detect malicious participants and protect privacy. In PID, we design a method to accumulate the offset of model parameters from participants in all current epochs to represent the moving tendency for model parameters. Thus, we can distinguish the adversaries from normal participants based on the accumulations in this incremental poisoning attack. We also use multiple trust domains to reduce the rate of misjudging benign participants as adversaries. Moreover, a differentiated differential privacy is utilized before the global model sending to protect the privacy of participants’ training datasets in PID. The experiments conducted on two real-world datasets under financial fraud detection scenario demonstrate that the PID reduces the fallout of adversaries detection (the rate of misjudging benign participants as adversaries) by at least 51.1% and improve the speed of detecting all malicious participants by at least 33.4% compared with two popular defense methods. Besides, the privacy preserving of PID is also effective.

DRL-empowered joint batch size and weighted aggregation adjustment mechanism for federated learning on non-IID data

To address the accuracy degradation as well as prolonged convergence time due to the inherent data heterogeneity among end-devices in federated learning (FL), we introduce the joint batch size and weighted aggregation adjustment problem, which is non-convex problem. To adjust optimal hyperparameters, we develop deep reinforcement learning (DRL) to empower a mechanism known as Batch size and Weighted aggregation Adjustment (BWA). Experimental evaluation demonstrates that BWA not only outperforms methods optimized solely from either a local training or server perspective but also achieves higher accuracy, with an increase of up to 5.53% compared to FedAvg, and additionally accelerates convergence speeds.

MRG-T: Mask-Relation-Guided Transformer for Remote Vision-Based Pedestrian Attribute Recognition in Aerial Imagery

Nowadays, with the rapid development of consumer Unmanned Aerial Vehicles (UAVs), utilizing UAV platforms for visual surveillance has become very attractive, and a key part of this is remote vision-based pedestrian attribute recognition. Pedestrian Attribute Recognition (PAR) is dedicated to predicting multiple attribute labels of a single pedestrian image extracted from surveillance videos and aerial imagery, which presents significant challenges in the computer vision community due to factors such as poor imaging quality and substantial pose variations. Despite recent studies demonstrating impressive advancements in utilizing complicated architectures and exploring relations, most of them may fail to fully and systematically consider the inter-region, inter-attribute, and region-attribute mapping relations simultaneously and be stuck in the dilemma of information redundancy, leading to the degradation of recognition accuracy. To address the issues, we construct a novel Mask-Relation-Guided Transformer (MRG-T) framework that consists of three relation modeling modules to fully exploit spatial and semantic relations in the model learning process. Specifically, we first propose a Masked Region Relation Module (MRRM) to focus on precise spatial attention regions to extract more robust features with masked random patch training. To explore the semantic association of attributes, we further present a Masked Attribute Relation Module (MARM) to extract intrinsic and semantic inter-attribute relations with an attribute label masking strategy. Based on the cross-attention mechanism, we finally design a Region and Attribute Mapping Module (RAMM) to learn the cross-modal alignment between spatial regions and semantic attributes. We conduct comprehensive experiments on three public benchmarks such as PETA, PA-100K, and RAPv1, and conduct inference on a large-scale airborne person dataset named PRAI-1581. The extensive experimental results demonstrate the superior performance of our method compared to state-of-the-art approaches and validate the effectiveness of mask-relation-guided modeling in the remote vision-based PAR task.

Neural Amortized Inference for Nested Multi-Agent Reasoning

Multi-agent interactions, such as communication, teaching, and bluffing, often rely on higher-order social inference, i.e., understanding how others infer oneself. Such intricate reasoning can be effectively modeled through nested multi-agent reasoning. Nonetheless, the computational complexity escalates exponentially with each level of reasoning, posing a significant challenge. However, humans effortlessly perform complex social inferences as part of their daily lives. To bridge the gap between human-like inference capabilities and computational limitations, we propose a novel approach: leveraging neural networks to amortize high-order social inference, thereby expediting nested multi-agent reasoning. We evaluate our method in two challenging multi-agent interaction domains. The experimental results demonstrate that our method is computationally efficient while exhibiting minimal degradation in accuracy.

Self-Distillation Regularized Connectionist Temporal Classification Loss for Text Recognition: A Simple Yet Effective Approach

Text recognition methods are gaining rapid development. Some advanced techniques, e.g., powerful modules, language models, and un- and semi-supervised learning schemes, consecutively push the performance on public benchmarks forward. However, the problem of how to better optimize a text recognition model from the perspective of loss functions is largely overlooked. CTC-based methods, widely used in practice due to their good balance between performance and inference speed, still grapple with accuracy degradation. This is because CTC loss emphasizes the optimization of the entire sequence target while neglecting to learn individual characters. We propose a self-distillation scheme for CTC-based model to address this issue. It incorporates a framewise regularization term in CTC loss to emphasize individual supervision, and leverages the maximizing-a-posteriori of latent alignment to solve the inconsistency problem that arises in distillation between CTC-based models. We refer to the regularized CTC loss as Distillation Connectionist Temporal Classification (DCTC) loss. DCTC loss is module-free, requiring no extra parameters, longer inference lag, or additional training data or phases. Extensive experiments on public benchmarks demonstrate that DCTC can boost text recognition model accuracy by up to 2.6%, without any of these drawbacks.

DataElixir: Purifying Poisoned Dataset to Mitigate Backdoor Attacks via Diffusion Models

Dataset sanitization is a widely adopted proactive defense against poisoning-based backdoor attacks, aimed at filtering out and removing poisoned samples from training datasets. However, existing methods have shown limited efficacy in countering the ever-evolving trigger functions, and often leading to considerable degradation of benign accuracy. In this paper, we propose DataElixir, a novel sanitization approach tailored to purify poisoned datasets. We leverage diffusion models to eliminate trigger features and restore benign features, thereby turning the poisoned samples into benign ones. Specifically, with multiple iterations of the forward and reverse process, we extract intermediary images and their predicted labels for each sample in the original dataset. Then, we identify anomalous samples in terms of the presence of label transition of the intermediary images, detect the target label by quantifying distribution discrepancy, select their purified images considering pixel and feature distance, and determine their ground-truth labels by training a benign model. Experiments conducted on 9 popular attacks demonstrates that DataElixir effectively mitigates various complex attacks while exerting minimal impact on benign accuracy, surpassing the performance of baseline defense methods.