让图注意力自适应学习稀疏程度，提升模型可解释性与效率。
When Should Graph Attention Be Sparse? Learning a Per-Edge Tsallis Index
- 通过可学习的Tsallis指数动态调节注意力稀疏性，支持从全局到边级的灵活控制。
- 学习得到的稀疏注意力能精准剪除42%错误连接，恢复后性能提升7.1点。
- 无需网格搜索，单次运行即可获得可解释的稀疏结构，适合追求高效与透明性的研究者。
图注意力通常使用softmax进行邻域得分归一化，这是在香农统计下熵最大的选择。然而，同质图和异质图需要不同的注意力分布形态，单一归一化无法兼顾两者。我们提出可学习的Tsallis图注意力（LTGA），其Tsallis熵指数 $q$ 与权重联合学习，可在重尾（$q<1$）、softmax（$q=1$）和紧支撑（$q>1$）之间连续插值，支持从全局标量到每边索引的四种粒度。在八个基准测试、十次种子实验中，LTGA-Edge平均排名最佳（2.75），但全样本检验不显著（$p=0.199$），且学习$q$未优于手动调参：验证集最优的固定网格达61.4%，α-entmax为62.2%，容量匹配的 $q o1$ 控制组为62.0%，而LTGA-Edge为61.7%。学习到的指数优势在于仅需一次运行替代网格搜索，并提供可解释机制：当 $q$ 偏离1时，42%的注意力系数被精确置零，这些边是选择性错误的，恢复它们带来7.1点增益，随机剪枝同率则损失13.0点。
原文摘要 · Abstract (English)
Graph attention normalizes neighborhood scores with softmax, the maximum-entropy choice under Shannon statistics. But homophilic and heterophilic graphs want different attention shapes, and one fixed normalization cannot serve both. We propose \textbf{LTGA} (\textbf{L}earnable \textbf{T}sallis \textbf{G}raph \textbf{A}ttention), a graph attention layer whose Tsallis entropic index $q$ is learned jointly with the weights, interpolating continuously between heavy-tailed ($q\!<\!1$), softmax ($q\!=\!1$) and compact-support ($q\!>\!1$) attention at four granularities from a global scalar to a per-edge index, under a bounded reparameterization that starts every model at the GAT baseline. Across eight benchmarks at ten seeds, LTGA-Edge takes the best average rank ($2.75$), but the omnibus test does not reject ($p\!=\!0.199$) and learning $q$ does not beat searching it: a validation-tuned frozen grid reaches $61.4\%$, tuned $α$-entmax $62.2\%$ and a capacity-matched $q\!\equiv\!1$ control $62.0\%$, against $61.7\%$ for LTGA-Edge. What the learned index buys is one run instead of a grid, and an interpretable mechanism: where $q$ leaves $1$, it prunes $42\%$ of attention coefficients to exactly zero, and those edges are selectively the wrong ones, restoring them costs $7.1$ points, while random pruning at the same rate costs $13.0$ more. Project page: https://kleyt0n.github.io/ltga
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。