Training language models and examining their linguistic behaviors have been a common protocol in computational linguistics for studying linguistic phenomena and modeling human language processing.However, work in this area is often limited to proof-of-concept demonstrations with arbitrary model configurations, without considering hyperparameter sensitivity, an important source of variation in model performance.In this work, we replicate three prior studies (Chang and Bergen, 2022; Hu et al., 2020b;Kuribayashi et al., 2024) with hyperparameters varied within a practical range, and show that modest hyperparameter changes can alter some qualitative conclusions about models' linguistic abilities and even reverse the ranking of model performance.Our results highlight the risk that prior work may have reflected optimization artifacts rather than the genuine inductive biases of model classes, and that hyperparameter sensitivity should receive more attention as a factor that can meaningfully influence model behavior.We suggest future work to report the variation of performance across the configuration space to enhance the reliability and generalizability of conclusions.

Single-channel EEG based sleep staging is of interest to researchers due to its broad application prospect in daily sleep monitoring recently. We proposed using contextual scalograms as input and developed a convolutional neural network with attention modules named Co-ScaleNet for sleep staging. The contextual scalograms were obtained by combining the same color channels of three original RGB scalograms from consecutive epochs, and a simple and efficient data augmentation was designed according to their various forms. The Co-ScaleNet consists of two main parts. Firstly, three parallel convolutional branches with attention modules correspondingly extract and fuse features from contextual scalograms at the top layers. The remaining part is a stack of lightweight blocks. We achieved an overall accuracy of 87.0% for healthy individuals, 84.7% for depressed patients. And we obtained comparable performance on the public Sleep-EDFx (82.8%), ISRUC (84.6%) and SHHS datasets (87.7%), including a high recall of N1. The contextual scalograms of R channel as input achieved the best performance, which conform to the features of interest in visual scoring. The attention modules improved the recall of N1 and N3. Overall, the contextual scalograms provided a novel scheme for both contextual information extraction and data augmentation. Our study successfully expanded its application to depression datasets, as well as patients with sleep apnea, demonstrating its wide applicability.

This letter describes a method of correcting ionospheric frequency modulation for a high-frequency hybrid sky-surface wave radar mounted on a shipborne platform. In the proposed method, azimuth-dependent sea clutter signals are first decomposed into monocomponent signals based on distinguishable differences in their directions of incidence. Afterward, based on the decomposed monocomponent signals, the statistical mean of the time derivatives of the signal phases, weighted by the signal amplitudes, is used to estimate the ionospheric frequency modulation. Finally, the estimated result is applied to the received data to compensate for the ionospheric contamination. Numerical results on simulated data demonstrate the effectiveness of the proposed algorithm.

It transpires that the irregularity in the structure of the ionospheric plasma plays a significant role on the ionospherically-propagated HF signals. In this paper, special attention has been paid to derive a simulator that can explicate the perturbed phase influence imposed by the ionosphere irregularities. This has been achieved by studying the space-time correlation as well as statistics of the perturbed phases so that the problem of perturbed phase simulation is recast as generating particular time series satisfying specific power spectrum and statistical distribution. Eventually, three scenarios corresponding to different irregularity fluctuation conditions are considered to verify the effectiveness of the proposed simulator. It is shown numerically that the obtained perturbed phases can agree well with the theoretical assumptions.

In high‐frequency (HF) hybrid sky‐surface wave radar, the first‐order sea clutter broadening is very severely under the influence of ionosphere and bistatic angle, which affects the detection of ship target. In this study, a cascaded method for ionospheric decontamination and sea clutter suppression to enable detection of ship target is presented. Firstly, the radar configuration, sea clutter broadening characteristics, and model are introduced based on the newly‐developed integrated HF sky‐surface wave experimental system. In the cascaded processing method, the time–frequency analysis method based on S 2 ‐method is used to correct the ionospheric phase contamination with large amplitude. Then, a novel forward–backward linear prediction algorithm is proposed to suppress the broadening bistatic sea clutter caused by bistatic angle, which is based on the distribution characteristics and the multi‐dimensional feature of first‐order sea clutter. Finally, the proposed method is examined by measured data. Experimental results indicate that the proposed method can well suppress the broadening sea clutter and improve the ship target detection performance.