Yong Hu

Yong Hu, Fandong Meng, Jie Zhou · tsinghua

In this paper, we present CSCD-NS, the first Chinese spelling check (CSC) dataset designed for native speakers, containing 40,000 samples from a Chinese social platform. Compared with existing CSC datasets aimed at Chinese learners, CSCD-NS is ten times larger in scale and exhibits a distinct error distribution, with a significantly higher proportion of word-level errors. To further enhance the data resource, we propose a novel method that simulates the input process through an input method, generating large-scale and high-quality pseudo data that closely resembles the actual error distribution and outperforms existing methods. Moreover, we investigate the performance of various models in this scenario, including large language models (LLMs), such as ChatGPT. The result indicates that generative models underperform BERT-like classification models due to strict length and pronunciation constraints. The high prevalence of word-level errors also makes CSC for native speakers challenging enough, leaving substantial room for improvement.

Cheng Kang, Jindich Prokop, Lei Tong et al.

This paper presents a new model architecture, gate with inhibition MLP (giMLP).The gate with inhibition on CycleMLP (gi-CycleMLP) can produce equal performance on the ImageNet classification task, and it also improves the BERT, Roberta, and DeBERTaV3 models depending on two novel techniques. The first is the gating MLP, where matrix multiplications between the MLP and the trunk Attention input in further adjust models' adaptation. The second is inhibition which inhibits or enhances the branch adjustment, and with the inhibition levels increasing, it offers models more muscular features restriction. We show that the giCycleMLP with a lower inhibition level can be competitive with the original CycleMLP in terms of ImageNet classification accuracy. In addition, we also show through a comprehensive empirical study that these techniques significantly improve the performance of fine-tuning NLU downstream tasks. As for the gate with inhibition MLPs on DeBERTa (giDeBERTa) fine-tuning, we find it can achieve appealing results on most parts of NLU tasks without any extra pretraining again. We also find that with the use of Gate With Inhibition, the activation function should have a short and smooth negative tail, with which the unimportant features or the features that hurt models can be moderately inhibited. The experiments on ImageNet and twelve language downstream tasks demonstrate the effectiveness of Gate With Inhibition, both for image classification and for enhancing the capacity of nature language fine-tuning without any extra pretraining.

Jiarui Zhang, Xiangyu Liu, Yong Hu et al.

Retrieval-Augmented Generation (RAG) significantly improves the performance of Large Language Models (LLMs) on knowledge-intensive tasks. However, varying response quality across LLMs under RAG necessitates intelligent routing mechanisms, which select the most suitable model for each query from multiple retrieval-augmented LLMs via a dedicated router model. We observe that external documents dynamically affect LLMs' ability to answer queries, while existing routing methods, which rely on static parametric knowledge representations, exhibit suboptimal performance in RAG scenarios. To address this, we formally define the new retrieval-augmented LLM routing problem, incorporating the influence of retrieved documents into the routing framework. We propose RAGRouter, a RAG-aware routing design, which leverages document embeddings and RAG capability embeddings with contrastive learning to capture knowledge representation shifts and enable informed routing decisions. Extensive experiments on diverse knowledge-intensive tasks and retrieval settings, covering open and closed-source LLMs, show that RAGRouter outperforms the best individual LLM and existing routing methods. With an extended score-threshold-based mechanism, it also achieves strong performance-efficiency trade-offs under low-latency constraints. The code and data are available at https://github.com/OwwO99/RAGRouter.

Yuyang Wang, Yukuan Hu, Xin Liu

Tensor product function (TPF) approximations have been widely adopted in solving high-dimensional problems, such as partial differential equations and eigenvalue problems, achieving desirable accuracy with computational overhead that scales linearly with problem dimensions. However, recent studies have underscored the extraordinarily high computational cost of TPFs on quantum many-body problems, even for systems with as few as three particles. A key distinction in these problems is the antisymmetry requirement on the unknown functions. In the present work, we rigorously establish that the minimum number of involved terms for a class of TPFs to be exactly antisymmetric increases exponentially fast with the problem dimension. This class encompasses both traditionally discretized TPFs and the recent ones parameterized by neural networks. Our proof exploits the link between the antisymmetric TPFs in this class and the corresponding antisymmetric tensors and focuses on the Canonical Polyadic rank of the latter. As a result, our findings reveal that low-rank TPFs are fundamentally unsuitable for high-dimensional problems where antisymmetry is essential.

Dong Chen, Yanzhe Wei, Zonglin He et al.

Large language models (LLMs) offer transformative potential for clinical decision support in spine surgery but pose significant risks through hallucinations, which are factually inconsistent or contextually misaligned outputs that may compromise patient safety. This study introduces a clinician-centered framework to quantify hallucination risks by evaluating diagnostic precision, recommendation quality, reasoning robustness, output coherence, and knowledge alignment. We assessed six leading LLMs across 30 expert-validated spinal cases. DeepSeek-R1 demonstrated superior overall performance (total score: 86.03 $\pm$ 2.08), particularly in high-stakes domains such as trauma and infection. A critical finding reveals that reasoning-enhanced model variants did not uniformly outperform standard counterparts: Claude-3.7-Sonnet's extended thinking mode underperformed relative to its standard version (80.79 $\pm$ 1.83 vs. 81.56 $\pm$ 1.92), indicating extended chain-of-thought reasoning alone is insufficient for clinical reliability. Multidimensional stress-testing exposed model-specific vulnerabilities, with recommendation quality degrading by 7.4% under amplified complexity. This decline contrasted with marginal improvements in rationality (+2.0%), readability (+1.7%) and diagnosis (+4.7%), highlighting a concerning divergence between perceived coherence and actionable guidance. Our findings advocate integrating interpretability mechanisms (e.g., reasoning chain visualization) into clinical workflows and establish a safety-aware validation framework for surgical LLM deployment.

Cheng Kang, Daniel Novak, Katerina Urbanova et al.

Large language models (LLMs) have demonstrated impressive generalization capabilities on specific tasks with human-written instruction data. However, the limited quantity, diversity, and professional expertise of such instruction data raise concerns about the performance of LLMs in psychotherapy tasks when provided with domain-specific instructions. To address this, we firstly propose Domain-Specific Assistant Instructions based on AlexanderStreet therapy, and secondly, we use an adaption fine-tuning method and retrieval augmented generation method to improve pre-trained LLMs. Through quantitative evaluation of linguistic quality using automatic and human evaluation, we observe that pre-trained LLMs on Psychotherapy Assistant Instructions outperform state-of-the-art LLMs response baselines. Our Assistant-Instruction approach offers a half-annotation method to align pre-trained LLMs with instructions and provide pre-trained LLMs with more psychotherapy knowledge.

Dong Chen, Huili Peng, Yong Hu et al.

Objective: To systematically quantify the effect of the camera view (frontal vs. lateral) on the accuracy of 2D markerless gait analysis relative to 3D motion capture ground truth. Methods: Gait data from 18 subjects were recorded simultaneously using frontal, lateral and 3D motion capture systems. Pose estimation used YOLOv8. Four metrics were assessed to evaluate agreement: Dynamic Time Warping (DTW) for temporal alignment, Maximum Cross-Correlation (MCC) for signal similarity, Kullback-Leibler Divergence (KLD) for distribution differences, and Information Entropy (IE) for complexity. Wilcoxon signed-rank tests (significance: $p < 0.05$) and Cliff's delta ($δ$) were used to measure statistical differences and effect sizes. Results: Lateral views significantly outperformed frontal views for sagittal plane kinematics: step length (DTW: $53.08 \pm 24.50$ vs. $69.87 \pm 25.36$, $p = 0.005$) and knee rotation (DTW: $106.46 \pm 38.57$ vs. $155.41 \pm 41.77$, $p = 0.004$). Frontal views were superior for symmetry parameters: trunk rotation (KLD: $0.09 \pm 0.06$ vs. $0.30 \pm 0.19$, $p < 0.001$) and wrist-to-hipmid distance (MCC: $105.77 \pm 29.72$ vs. $75.20 \pm 20.38$, $p = 0.003$). Effect sizes were medium-to-large ($δ: 0.34$--$0.76$). Conclusion: Camera view critically impacts gait parameter accuracy. Lateral views are optimal for sagittal kinematics; frontal views excel for trunk symmetry. Significance: This first systematic evidence enables data-driven camera deployment in 2D gait analysis, enhancing clinical utility. Future implementations should leverage both views via disease-oriented setups.

Zhihao Liu, Jian Lou, Yuke Hu et al.

Large language models (LLMs) are trained on massive datasets that may include private or copyrighted content. Due to growing privacy and ownership concerns, data owners may request the removal of their data from trained models. Machine unlearning provides a practical solution by removing the influence of specific data without full retraining. However, most existing methods lack a sound forgetting boundary, causing some samples to be under-forgotten, leaving residual leakage risks, while others remain over-forgotten at the expense of degraded utility. In this work, we propose EAGLE-PC (Entanglement-Awareness Guided Loss Reweighting with Proxy Constraint), a novel unlearning framework that addresses these limitations through two key components. First, entanglement-awareness guided loss reweighting determines the forgetting effort of each sample by measuring its similarity to retain samples in the embedding space, enabling more targeted and effective unlearning. Second, a proxy constraint leveraging ICL (In-Context Learning) generated test data softly regularizes the forgetting process, effectively mitigating over-forgetting. EAGLE-PC is compatible with existing gradient-based objectives and serves as a plug-and-play enhancement. We evaluate EAGLE-PC on the TOFU and MUSE benchmarks, showing consistent improvements in the forgetting-utility trade-off across multiple LLMs. Combined with the NPO+GD optimizer, it approaches full retraining performance, offering a scalable and robust unlearning solution.

Yuxin Liu, Zhenxi Song, Guoyang Xu et al.

Brain-computer interface (BCI) based on steady-state visual evoked potentials (SSVEP) is a popular paradigm for its simplicity and high information transfer rate (ITR). Accurate and fast SSVEP decoding is crucial for reliable BCI performance. However, conventional decoding methods demand longer time windows, and deep learning models typically require subject-specific fine-tuning, leaving challenges in achieving optimal performance in cross-subject settings. This paper proposed a biofocal masking attention-based method (SSVEP-BiMA) that synergistically leverages the native and symmetric-antisymmetric components for decoding SSVEP. By utilizing multiple signal representations, the network is able to integrate features from a wider range of sample perspectives, leading to more generalized and comprehensive feature learning, which enhances both prediction accuracy and robustness. We performed experiments on two public datasets, and the results demonstrate that our proposed method surpasses baseline approaches in both accuracy and ITR. We believe that this work will contribute to the development of more efficient SSVEP-based BCI systems.

Wen Yu, Baiying Lei, Yanyan Shen et al.

The diagnosis of early stages of Alzheimer's disease (AD) is essential for timely treatment to slow further deterioration. Visualizing the morphological features for the early stages of AD is of great clinical value. In this work, a novel Multidirectional Perception Generative Adversarial Network (MP-GAN) is proposed to visualize the morphological features indicating the severity of AD for patients of different stages. Specifically, by introducing a novel multidirectional mapping mechanism into the model, the proposed MP-GAN can capture the salient global features efficiently. Thus, by utilizing the class-discriminative map from the generator, the proposed model can clearly delineate the subtle lesions via MR image transformations between the source domain and the pre-defined target domain. Besides, by integrating the adversarial loss, classification loss, cycle consistency loss and \emph{L}1 penalty, a single generator in MP-GAN can learn the class-discriminative maps for multiple-classes. Extensive experimental results on Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset demonstrate that MP-GAN achieves superior performance compared with the existing methods. The lesions visualized by MP-GAN are also consistent with what clinicians observe.