1
Hierarchical Attention via Domain Decomposition
高效近似注意力机制新突破:用域分解实现层次化注意力,降低计算复杂度。
arXiv:2606.18525v1 Announce Type: new Abstract: We propose a hierarchical attention mechanism based on two-level overlapping Schwarz domain decomposit…
高效近似注意力机制新突破:用域分解实现层次化注意力,降低计算复杂度。
arXiv:2606.18525v1 Announce Type: new Abstract: We propose a hierarchical attention mechanism based on two-level overlapping Schwarz domain decomposit…
提出可微分自适应稀疏层次注意力机制,显著提升长序列建模效率与计算可扩展性
arXiv:2605.18753v1 Announce Type: cross Abstract: Current hierarchical attention methods, such as NSA and InfLLMv2, select the top-k relevant key-valu…