Distributed optimization is pivotal for large-scale signal processing and machine learning, yet communication overhead remains a major bottleneck. Low-rank gradient com-pression, in which the transmitted gradients are approximated by low-rank matrices to reduce communication, offers a promising remedy. Existing methods typically adopt either randomized or greedy compression strategies: randomized approaches project gradients onto randomly chosen subspaces, introducing high variance and degrading empirical performance; greedy methods select the most informative subspaces, achieving strong empirical results but lacking convergence guarantees. To address this gap, we propose GreedyLore-the first Greedy Low-Rank gradient compression algorithm for distributedlearning with rigorous convergence guarantees. GreedyLore incorporates error feedback to correct the bias introduced by greedy compression and introduces a semi-lazy subspace update that ensures the com-pression operator remains contractive throughout all iterations. With these techniques, we prove that GreedyLore achieves a convergence rate of$\mathcal{O}(\sigma/\sqrt{NT}^{-}+1/T)$under standard optimizers such as Adam-marking the first linear speedup convergence rate for low-rank gradient compression. Extensive experiments are conducted to validate our theoretical findings.

Subspace optimization algorithms, such as GaLore (Zhao et al., 2024), have gained attention for pre-training and fine-tuning large language models (LLMs) due to their memory efficiency. However, their convergence guarantees remain unclear, particularly in stochastic settings. In this paper, we reveal that GaLore does not always converge to the optimal solution and provide an explicit counterexample to support this finding. We further explore the conditions under which GaLore achieves convergence, showing that it does so when either (i) a sufficiently large mini-batch size is used or (ii) the gradient noise is isotropic. More significantly, we introduce GoLore (Gradient random Low-rank projection), a novel variant of GaLore that provably converges in typical stochastic settings, even with standard batch sizes. Our convergence analysis extends naturally to other subspace optimization algorithms. Finally, we empirically validate our theoretical results and thoroughly test the proposed mechanisms. Codes are available at https://github.com/pkumelon/Golore.

Electroencephalogram (EEG) signals, as a reliable biological indicator, have been widely used in fatigue driving detection due to their capacity to reflect a driver's cognitive and neural response state. However, EEG signals have problems such as imbalanced data distribution, significant differences between subjects, and complex scenes, which affect the detection effect. Small commonalities between input objects can be interpreted as important information about an entire sample. Therefore, to retain as much information as possible, We design a new approach for integrating fuzzy features, comprehensive adaptive interpretable TSK fuzzy classifier(CAI-TSK-FC). It not only captures the features of multiple subclassifiers more efficiently and alleviates the dataset imbalance problem. Also, it can reduce the accumulation of error information by randomly retaining fuzzy rules as well as normalization. Finally, we linearly combine the results of multiple subclassifiers to comprehensively consider the learning effect of multiple subclassifiers to adapt to different subjects and datasets. Experiments conducted on both self-made and public datasets (SEED-VIG) show that CAI-TSK-FC has good performance and interpretability on different EEG fatigue driving datasets. In comparison to existing methods, it achieves an accuracy improvement of 3.15% and 1.52%, respectively, as well as a specificity improvement of 4.72% and 0.91%, respectively.

Neuroscience research indicates that the interaction among different functional regions of the brain plays a crucial role in driving various cognitive tasks. Existing studies have primarily focused on constructing either local or global functional connectivity maps within the brain, often lacking an adaptive approach to fuse functional brain regions and explore latent relationships between localization during different cognitive tasks. This paper introduces a novel approach called the Local-Ascending-Global Learning Strategy (LAG) to uncover higher-level latent topological patterns among functional brain regions. The strategy initiates from the local connectivity of individual brain functional regions and develops a K-Level Self-Adaptive Ascending Network (SALK) to dynamically capture strong connectivity patterns among brain regions during different cognitive tasks. Through the step-by-step fusion of brain regions, this approach captures higher-level latent patterns, shedding light on the progressively adaptive fusion of various brain functional regions under different cognitive tasks. Notably, this study represents the first exploration of higher-level latent patterns through progressively adaptive fusion of diverse brain functional regions under different cognitive tasks. The proposed LAG strategy is validated using datasets related to fatigue (SEED-VIG), emotion (SEED-IV), and motor imagery (BCI_C_IV_2a). The results demonstrate the generalizability of LAG, achieving satisfactory outcomes in independent-subject experiments across all three datasets. This suggests that LAG effectively characterizes higher-level latent patterns associated with different cognitive tasks, presenting a novel approach to understanding brain patterns in varying cognitive contexts.

Electroencephalography (EEG) signal has been recognized as an effective fatigue detection method, which can intuitively reflect the drivers' mental state. However, the research on multi-dimensional features in existing work could be much better. The instability and complexity of EEG signals will increase the difficulty of extracting data features. More importantly, most current work only treats deep learning models as classifiers. They ignored the features of different subjects learned by the model. Aiming at the above problems, this paper proposes a novel multi-dimensional feature fusion network, CSF-GTNet, based on time and space-frequency domains for fatigue detection. Specifically, it comprises Gaussian Time Domain Network (GTNet) and Pure Convolutional Spatial Frequency Domain Network (CSFNet). The experimental results show that the proposed method effectively distinguishes between alert and fatigue states. The accuracy rates are 85.16% and 81.48% on the self-made and SEED-VIG datasets, respectively, which are higher than the state-of-the-art methods. Moreover, we analyze the contribution of each brain region for fatigue detection through the brain topology map. In addition, we explore the changing trend of each frequency band and the significance between different subjects in the alert state and fatigue state through the heat map. Our research can provide new ideas in brain fatigue research and play a specific role in promoting the development of this field.

Studying brain activity and deciphering the information in electroencephalogram (EEG) signals has become an emerging research field, and substantial advances have been made in the EEG-based classification of emotions. However, using different EEG features and complementarity to discriminate other emotions is still challenging. Most existing models extract a single temporal feature from the EEG signal while ignoring the crucial temporal dynamic information, which, to a certain extent, constrains the classification capability of the model. To address this issue, we propose an Attention-Based Depthwise Parameterized Convolutional Gated Recurrent Unit (AB-DPCGRU) model and validate it with the mixed experiment on the SEED and SEED-IV datasets. The experimental outcomes revealed that the accuracy of the model outperforms the existing state-of-the-art methods, which confirmed the superiority of our approach over currently popular emotion recognition models.