AI 中文总结
本研究提出即插即用模块ZIPBrain，利用EEG低SNR特性减少token数量，使EEG基础模型推理速度提升32.7%（最高41.8%）且平均准确率提升1.3%-10.5%，可本地部署。
AI 中文摘要
本研究探讨能否在不损失准确率的前提下，使脑电（EEG）基础模型（EFMs）变得更快且可本地部署。EEG基础模型是提供通用强表征的主流趋势，但它们的计算负担随输入长度呈二次增长，阻碍了在资源受限场景（尤其是实时临床监测）中的部署。EEG的低信噪比（SNR）进一步表明，许多token是冗余的，可在几乎不损失准确率的情况下进行压缩。我们提出ZIPBrain，一种新型的感知EEG token池化模块，利用低SNR特性减少token数量。给定token序列，ZIPBrain将token划分为冗余组和独特组，再将每个冗余token与其在独特组中最相似的对应token合并。此外，ZIPBrain是无需训练的即插即用模块，可无缝集成到标准Transformer编码器中，计算开销可忽略不计。对多个EEG基础模型进行的大量实验表明，ZIPBrain具有强通用性，相比基线模型实现了1.3%-10.5%的平均提升，同时与原始EEG基础模型相比，将挂钟推理时间减少了32.7%（使用CUDA Graph时最多减少41.8%）。
英文摘要
This work investigates whether Electroencephalograph (EEG) foundation models (EFMs) can be made faster and locally deployable without sacrificing accuracy. EEG foundation models are a major trend, offering strong general-purpose representations. However, their computational burden grows quadratically with input length, hindering deployment on resource-constrained scenario, particularly for real-time clinical monitoring. EEG's low SNR further suggests many of these tokens are redundant and compressible with little accuracy cost. We propose ZIPBrain, a novel redundancy-aware EEG token pooling module that leverages this low-SNR characteristic to reduce token count. Given a token sequence, ZIPBrain partitions tokens into redundant and unique groups, then merges each redundant token with its most similar counterpart in the unique group. Furthermore, ZIPBrain serves as a training-free, plug-and-play module that seamlessly integrates into standard Transformer encoders with negligible computational overhead. Extensive experiments across multiple EEG foundation models show ZIPBrain's strong versatility, achieving 1.3%-10.5% average improvement over baselines, while reducing wall-clock inference time by 32.7% (up to 41.8% with CUDA Graph) compared to the original EEG foundation models.
Comments7 pages(14 pages including appendix), 5 figures
