Spatio-temporal data prediction is a fundamental task in urban computing, benefiting a variety of real-life applications such as traffic forecasting and environmental monitoring. Due to the dynamic and time-involving nature of spatio-temporal data, researchers have increasingly emphasized online prediction. However, existing approaches (e.g., URCL) typically rely on data-replay strategies, which require storing large volumes of historical data to frequently update their models with new inputs. These methods impose substantial costs, including frequent buffer construction, high storage requirements, and increased training complexity. Furthermore, the single-pass nature of online data, combined with the constrained resources of online environments, highlights the urgent need for more efficient and lightweight solutions for online spatio-temporal prediction. To address these challenges, we propose Storm, a knowledge distillation driven continual learning framework. Storm introduces Dynamic Knowledge Distillation (DKD), leveraging an ever-evolving teacher model to train an effective student model. To optimize efficiency, Storm employs a Mixture-of-Experts (MoE) mechanism, which dynamically switches between the original training mode and the DKD mode. This hybrid design enables low-cost online learning while addressing the stabilityplasticity dilemma. To fully leverage single-pass online data, Storm integrates effective data augmentation methods tailored to the dynamic nature of spatio-temporal data. Moreover, Storm incorporates a Gradual Parameter Freezing (GPF) module to progressively reduce computational costs during online training. Extensive experiments conducted on four real-world datasets, evaluated across short-term, medium-term, and long-term prediction horizons, demonstrate the superiority of Storm. Specifically, Storm: (i) provides a general online training extension for various offline spatio-temporal models, and (ii) achieves remarkable improvements, e.g., up to 14.24% accuracy gains while requiring only 0.3% of the training and inference time compared to the state-of-the-art URCL framework. The source code is publicly available at https://github.com/ZJU-DAILY/Storm.

Federated learning (FL) has been regarded as a promising paradigm for enabling distributed model training over resource-limited edge devices. Although FL maintains data locality and enhances model generalization, it faces challenges such as model leakage and pressure from frequent model updates. Some existing schemes, such as differential privacy and model encryption, can partially alleviate these issues while sacrificing the accuracy of the modeling training or increasing the computational overheads in training. To address this issue, we design a low-complexity linear collaborative FL (LCFL) framework to enhance the privacy and accuracy of FL. Specifically, we propose the collaborative secrecy transmission (CST) algorithm by integrating a variant of Shamir's secret-sharing with the model segmentation, which can compresses and encrypts the local models for FL. The decoding complexity of the CST algorithm is only$O(N^{3})$under the compression ratio of$N$, which reduces the communication overhead and computational complexity. We conduct a quantitative analysis of the model error induced by the CST algorithm and derive its closed-form upper bound. Within LCFL, we formulate an optimization problem to maximize the global model accuracy in wireless FL by optimizing compression ratios, bandwidth allocation, and transmit-powers. Subsequently, we propose a low-complexity algorithm to solve this problem effectively. Numerical simulations demonstrate the efficacy of LCFL in improving FL's accuracy and security, and the results validate the efficiency of the proposed optimization scheme for wireless FL. The source code can be downloaded from the Github:https://github.com/MinITerence/LCFL.

Electrocardiograms (ECG), essential for diagnosing cardiovascular diseases. In recent years, the application of deep learning techniques has significantly improved the performance of ECG signal classification. Multi-resolution feature analysis, which captures and processes information at different time scales, can extract subtle changes and overall trends in ECG signals, showing unique advantages. However, common multi-resolution analysis methods based on simple feature addition or concatenation may lead to the neglect of low-resolution features, affecting model performance. To address this issue, this paper proposes the Multi-Resolution Mutual Learning Network (MRMNet). MRM-Net includes a dual-resolution attention architecture and a feature complementary mechanism. The dual-resolution attention architecture processes high-resolution and low-resolution features in parallel. Through the attention mechanism, the high-resolution and low-resolution branches can focus on subtle waveform changes and overall rhythm patterns, enhancing the ability to capture critical features in ECG signals. Meanwhile, the feature complementary mechanism introduces mutual feature learning after each layer of the feature extractor. This allows features at different resolutions to reinforce each other, thereby reducing information loss and improving model performance and robustness. Experiments on the PTB-XL and CPSC2018 datasets demonstrate that MRM-Net significantly outperforms existing methods in multi-label ECG classification performance. The code for our framework will be publicly available at https://github.com/wxhdf/MRM.

Extracting invariant representations in unlabeled electrocardiogram (ECG) signals is a challenge for deep neural networks (DNNs). Contrastive learning is a promising method for unsupervised learning. However, it should improve its robustness to noise and learn the spatiotemporal and semantic representations of categories, just like cardiologists. This article proposes a patient-level adversarial spatiotemporal contrastive learning (ASTCL) framework, which includes ECG augmentations, an adversarial module, and a spatiotemporal contrastive module. Based on the ECG noise attributes, two distinct but effective ECG augmentations, ECG noise enhancement, and ECG noise denoising, are introduced. These methods are beneficial for ASTCL to enhance the robustness of the DNN to noise. This article proposes a self-supervised task to increase the antiperturbation ability. This task is represented as a game between the discriminator and encoder in the adversarial module, which pulls the extracted representations into the shared distribution between the positive pairs to discard the perturbation representations and learn the invariant representations. The spatiotemporal contrastive module combines spatiotemporal prediction and patient discrimination to learn the spatiotemporal and semantic representations of categories. To learn category representations effectively, this article only uses patient-level positive pairs and alternately uses the predictor and the stop-gradient to avoid model collapse. To verify the effectiveness of the proposed method, various groups of experiments are conducted on four ECG benchmark datasets and one clinical dataset compared with the state-of-the-art methods. Experimental results showed that the proposed method outperforms the state-of-the-art methods.

Visual curve completion is a fundamental problem in understanding the principles of the human visual system. This problem is usually divided into two problems: a grouping problem and a shape problem. On one hand, though perception of the visually completed curve is clearly a global task (for example, a human perceives the Kanizsa triangle only when seeing all three black objects), conventional methods for solving the grouping problem are generally based on local Gestalt laws. On the other hand, the shape of the visually completed curve is usually recovered by minimizing shape energy in existing methods. However, not only do these methods lack mechanisms to adjust the shape of the recovered visual curve using perceptual, psychophysical, and neurophysiological knowledge, but it is also difficult to calculate an explicit representation of the visually completed curve. In this paper, we present a systematic computational model for generating a visually completed curve. Firstly, based on recent studies of perception, psychophysics, and neurophysiology, we formulate a grouping procedure based on the human visual system by seeking a minimum Hamiltonian cycle in a graph, solving the grouping problem in a global manner. Secondly, we employ a Bézier curve-based model to represent the visually completed curve. Not only is an explicit representation deduced, but we also present a means to integrate knowledge from related areas, such as perception, psychophysics, and neurophysiology, and so on. The proposed computational model has been validated using many modal and amodal completion examples, and desirable results were obtained.