arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2302.00049cs.LG
收藏

Transformer遇上定向图

Transformers Meet Directed Graphs

Technical University of Munich(慕尼黑工业大学);Google DeepMind(谷歌DeepMind)

Simon Geisler, Yujia Li, Daniel Mankowitz, Ali Taylan Cemgil, Stephan Günnemann, Cosmin Paduraru

AI总结:

本文针对定向图提出两种方向与结构感知的位置编码方法,并在排序网络测试和源代码理解等任务上验证其有效性,在Code2基准上相对提升14.7%。

摘要:

Transformer最初被提出作为文本的序列到序列模型,但已成为包括图像、音频、视频和无向图在内的多种模态的关键技术。然而,针对定向图的Transformer却是一个令人惊讶的未被充分探索的课题,尽管它们适用于包括源代码和逻辑电路在内的普遍领域。在这项工作中,我们提出了两种针对定向图的方向和结构感知位置编码:(1)磁拉普拉斯算子的特征向量——组合拉普拉斯算子的方向感知推广;(2)定向随机游走编码。实验上,我们表明额外的方向性信息在各种下游任务中是有用的,包括排序网络的正确性测试和源代码理解。结合以数据流为中心的图构建,我们的模型在Open Graph Benchmark Code2上相对先前最先进水平提升了14.7%。

英文摘要:

Transformers were originally proposed as a sequence-to-sequence model for text but have become vital for a wide range of modalities, including images, audio, video, and undirected graphs. However, transformers for directed graphs are a surprisingly underexplored topic, despite their applicability to ubiquitous domains, including source code and logic circuits. In this work, we propose two direction- and structure-aware positional encodings for directed graphs: (1) the eigenvectors of the Magnetic Laplacian - a direction-aware generalization of the combinatorial Laplacian; (2) directional random walk encodings. Empirically, we show that the extra directionality information is useful in various downstream tasks, including correctness testing of sorting networks and source code understanding. Together with a data-flow-centric graph construction, our model outperforms the prior state of the art on the Open Graph Benchmark Code2 relatively by 14.7%.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑