With the widespread deployment of large models on edge devices, more and more SpatioTemporal (ST) data processing tasks need to be performed locally on these devices. However, edge devices often face challenges, such as limited computational resources, stringent real-time and robustness requirements, scarcity of labeled data, and difficulties with manual intervention. To address these challenges, we propose a lightweight, fast, and robust method, named$\ell _{0}$-norm Rank-Adaptive Spatiotemporal data Completion Neural Network ($\ell _{0}$-RASC-NN). We first formulate the target as an optimization model and employ Block Coordinate Descent (BCD) combined with alternating optimization for iterative solving. This iterative process is then transformed into an artificial neural network using a deep unrolling algorithm. This approach not only reduces computational cost and improves real-time performance but also better addresses the challenge of limited labeled data compared to conventional deep learning methods. To enhance robustness, the model incorporates an$\ell _{0}$-norm regularization term and a dynamic threshold adjustment strategy to handle anomalies. Furthermore, an adaptive rank selection mechanism and hyperparameter reparameterization are introduced to minimize the need for manual intervention. Experiments on five real-world ST datasets demonstrate that our method outperforms other state-of-the-art approaches in both ST data recovery and anomaly detection. Notably, under the same conditions,$\ell _{0}$-RASC-NN reduces computational time by one to two orders of magnitude compared to existing methods, while maintaining or even enhancing recovery accuracy. The code of our proposed method is provided athttps://tinyurl.com/36crh4v9.

Machine intelligence fault prediction (MIFP) is crucial for ensuring complex systems’ safe and reliable operation. While deep learning has become the mainstream tool for MIFP due to its excellent learning abilities, its interpretability is limited, and it struggles to learn frequencies, making it challenging to understand the physical knowledge of signals at the frequency level. Therefore, this article proposes a physically interpretable wavelet-guided network (WaveGNet) with deep frequency separation for MIFP, inspired by the sound theoretical basis and physical meaning of discrete wavelet transform (DWT). WaveGNet expands the feature learning space of CNN into the frequency domain, allowing for a better understanding of the physical insights behind the frequency level. Specifically, WaveGNet involves a derivable and learnable frequency learning layer (FL-Layer) consisting of a wavelet-driven frequency decomposition module and a convolution-driven feature learning module. Multiple DWT-driven FL-Layers are used in WaveGNet to achieve deep frequency decomposition and multiresolution frequency feature learning in a coarse-to-fine manner. The effectiveness of WaveGNet was evaluated in real high-speed train wheel wear monitoring and high-speed aviation bearing fault diagnosis cases. Experimental results showed that WaveGNet outperforms cutting-edge deep learning algorithms and has excellent fault diagnosis and prediction abilities. Furthermore, an in-depth analysis of the learning mechanism of wavelet-driven CNN from the frequency domain perspective was conducted.

Nowadays, profile data mining techniques facilitate effective process monitoring, quality control, fault diagnosis, etc., with considerable benefits to manufacturing industry. However, regarding the complex system in modern manufacturing industry, there are two significant challenges for application development based on profile data mining. First, the staggering data volume leads to high memory and computational requirements. Second, the noisy signals in collected data may deteriorate useful information and model performance. This article proposes a novel algorithm for profile data mining called profile abstract, which simultaneously enables profile data compression and segmentation. The proposed algorithm mainly considers the scenario of fault diagnosis and can be utilized as a pre-processing step to address the above challenges. Profile abstract seeks to find a subset of raw data or a group of models as representatives that preserve the essential characteristics of raw data. Finding the data representatives helps reduce data redundancy while maintaining the model performance. Model representatives assist in describing the complex pattern of the profile, which can be used for pattern-based data segmentation. After data segmentation, information gain is adopted to determine the critical primitives for model improvement. In this article, validation of the proposed method's superiority is performed with two datasets from a real production line and one simulation dataset.

Short-term forecasting of passenger flow is critical for transit management and crowd regulation. Spatial dependencies, temporal dependencies, inter-station correlations driven by other latent factors, and exogenous factors bring challenges to the short-term forecasts of passenger flow of urban rail transit networks. An innovative deep learning approach, Multi-Graph Convolutional-Recurrent Neural Network (MGC-RNN) is proposed to forecast passenger flow in urban rail transit systems to incorporate these complex factors. We propose to use multiple graphs to encode the spatial and other heterogenous inter-station correlations. The temporal dynamics of the inter-station correlations are also modeled via the proposed multi-graph convolutional-recurrent neural network structure. Inflow and outflow of all stations can be collectively predicted with multiple time steps ahead via a sequence to sequence(seq2seq) architecture. The proposed method is applied to the short-term forecasts of passenger flow in Shenzhen Metro, China. The experimental results show that MGC-RNN outperforms the benchmark algorithms in terms of forecasting accuracy. Besides, it is found that the inter-station driven by network distance, network structure, and recent flow patterns are significant factors for passenger flow forecasting. Moreover, the architecture of LSTM-encoder-decoder can capture the temporal dependencies well. In general, the proposed framework could provide multiple views of passenger flow dynamics for fine prediction and exhibit a possibility for multi-source heterogeneous data fusion in the spatiotemporal forecast tasks.

In the above article[1],Table I,III, andIVshould show “N/m” instead of “kN/m” and they should also show “Ns/m” instead of “kNs/m.” The revised tables are shown below.

A novel data-driven framework to monitor the health status of high-speed rail suspension system by measuring train vibrations is proposed herein. Unlike existing methods, this framework does not rely on sophisticated dynamic models or high-fidelity simulations; it combines the power of data and domain knowledge to generate a model that can be trained quickly and adapted easily to different rail systems. In addition, the framework includes a module to generate a training dataset, tackling a typical challenge in real-world system monitoring, namely, the lack of labeled data due to practical limits. Based on the multi-output support vector regression (MSVR), the proposed framework can monitor the stiffness and damping coefficients of the suspension system using vibration signals measured on trains in real time. The framework comprises three modules. First, a simple suspension system dynamics model is built to generate a training dataset. Furthermore, key features are extracted from frequency response curves to reflect the impact of spring and damper degradation. Subsequently, a supervised learning model based on the MSVR is built to predict the stiffness and damping coefficients of suspension systems from features extracted in the second module. Once the model is built, real-time monitoring can be achieved by feeding the vibration signals as they are collected during operations. The proposed framework was evaluated on simulation data for its accuracy and tested on real-world operational data for its practicability.

In many real-world classification problems, accurate prediction of membership probabilities is critical for further decision making. The probability calibration problem studies how to map scores obtained from one classification algorithm to membership probabilities. The requirement of non-decreasingness for this mapping involves an infinite number of inequality constraints, which makes its estimation computationally intractable. For the sake of this difficulty, existing methods failed to achieve four desiderata of probability calibration: universal flexibility, non-decreasingness, continuousness and computational tractability. This paper proposes a method with shape-restricted polynomial regression, which satisfies all four desiderata. In the method, the calibrating function is approximated with monotone polynomials, and the continuously-constrained requirement of monotonicity is equivalent to some semidefinite constraints. Thus, the calibration problem can be solved with tractable semidefinite programs. This estimator is both strongly and weakly universally consistent under a trivial condition. Experimental results on both artificial and real data sets clearly show that the method can greatly improve calibrating performance in terms of reliability-curve related measures.