Pengyun Wang, Yadi Cao, Chris Russell, Siyu Heng, Junyu Luo, Yanxin Shen, Xiao Luo
{"title":"DELTA:带拓扑不确定性的双一致性迭代,用于主动图域自适应","authors":"Pengyun Wang, Yadi Cao, Chris Russell, Siyu Heng, Junyu Luo, Yanxin Shen, Xiao Luo","doi":"arxiv-2409.08946","DOIUrl":null,"url":null,"abstract":"Graph domain adaptation has recently enabled knowledge transfer across\ndifferent graphs. However, without the semantic information on target graphs,\nthe performance on target graphs is still far from satisfactory. To address the\nissue, we study the problem of active graph domain adaptation, which selects a\nsmall quantitative of informative nodes on the target graph for extra\nannotation. This problem is highly challenging due to the complicated\ntopological relationships and the distribution discrepancy across graphs. In\nthis paper, we propose a novel approach named Dual Consistency Delving with\nTopological Uncertainty (DELTA) for active graph domain adaptation. Our DELTA\nconsists of an edge-oriented graph subnetwork and a path-oriented graph\nsubnetwork, which can explore topological semantics from complementary\nperspectives. In particular, our edge-oriented graph subnetwork utilizes the\nmessage passing mechanism to learn neighborhood information, while our\npath-oriented graph subnetwork explores high-order relationships from\nsubstructures. To jointly learn from two subnetworks, we roughly select\ninformative candidate nodes with the consideration of consistency across two\nsubnetworks. Then, we aggregate local semantics from its K-hop subgraph based\non node degrees for topological uncertainty estimation. To overcome potential\ndistribution shifts, we compare target nodes and their corresponding source\nnodes for discrepancy scores as an additional component for fine selection.\nExtensive experiments on benchmark datasets demonstrate that DELTA outperforms\nvarious state-of-the-art approaches.","PeriodicalId":501032,"journal":{"name":"arXiv - CS - Social and Information Networks","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2024-09-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"DELTA: Dual Consistency Delving with Topological Uncertainty for Active Graph Domain Adaptation\",\"authors\":\"Pengyun Wang, Yadi Cao, Chris Russell, Siyu Heng, Junyu Luo, Yanxin Shen, Xiao Luo\",\"doi\":\"arxiv-2409.08946\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Graph domain adaptation has recently enabled knowledge transfer across\\ndifferent graphs. However, without the semantic information on target graphs,\\nthe performance on target graphs is still far from satisfactory. To address the\\nissue, we study the problem of active graph domain adaptation, which selects a\\nsmall quantitative of informative nodes on the target graph for extra\\nannotation. This problem is highly challenging due to the complicated\\ntopological relationships and the distribution discrepancy across graphs. In\\nthis paper, we propose a novel approach named Dual Consistency Delving with\\nTopological Uncertainty (DELTA) for active graph domain adaptation. Our DELTA\\nconsists of an edge-oriented graph subnetwork and a path-oriented graph\\nsubnetwork, which can explore topological semantics from complementary\\nperspectives. In particular, our edge-oriented graph subnetwork utilizes the\\nmessage passing mechanism to learn neighborhood information, while our\\npath-oriented graph subnetwork explores high-order relationships from\\nsubstructures. To jointly learn from two subnetworks, we roughly select\\ninformative candidate nodes with the consideration of consistency across two\\nsubnetworks. Then, we aggregate local semantics from its K-hop subgraph based\\non node degrees for topological uncertainty estimation. To overcome potential\\ndistribution shifts, we compare target nodes and their corresponding source\\nnodes for discrepancy scores as an additional component for fine selection.\\nExtensive experiments on benchmark datasets demonstrate that DELTA outperforms\\nvarious state-of-the-art approaches.\",\"PeriodicalId\":501032,\"journal\":{\"name\":\"arXiv - CS - Social and Information Networks\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-13\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - CS - Social and Information Networks\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.08946\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - CS - Social and Information Networks","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.08946","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
最近,图域适应技术实现了跨不同图的知识转移。然而,如果没有目标图的语义信息,在目标图上的性能仍然远远不能令人满意。为了解决这个问题,我们研究了主动图域适配问题,即选择目标图上的少量信息节点进行额外标注。由于图之间存在复杂的拓扑关系和分布差异,这个问题极具挑战性。在本文中,我们提出了一种名为 "带拓扑不确定性的双一致性掘取(Dual Consistency Delving withTopological Uncertainty,DELTA)"的主动图域适应新方法。我们的 DELTA 由面向边缘的图子网络和面向路径的图子网络组成,可以从互补的角度探索拓扑语义。其中,面向边缘的图子网络利用消息传递机制学习邻域信息,而面向路径的图子网络则从子结构中探索高阶关系。为了从两个子网络中联合学习,我们在考虑两个子网络一致性的基础上,粗略地选择有信息的候选节点。然后,我们根据节点度从其 K 跳子图中汇总局部语义,以进行拓扑不确定性估计。为了克服潜在的分布偏移,我们比较了目标节点和其相应来源节点的差异分数,作为精细选择的额外组成部分。
DELTA: Dual Consistency Delving with Topological Uncertainty for Active Graph Domain Adaptation
Graph domain adaptation has recently enabled knowledge transfer across
different graphs. However, without the semantic information on target graphs,
the performance on target graphs is still far from satisfactory. To address the
issue, we study the problem of active graph domain adaptation, which selects a
small quantitative of informative nodes on the target graph for extra
annotation. This problem is highly challenging due to the complicated
topological relationships and the distribution discrepancy across graphs. In
this paper, we propose a novel approach named Dual Consistency Delving with
Topological Uncertainty (DELTA) for active graph domain adaptation. Our DELTA
consists of an edge-oriented graph subnetwork and a path-oriented graph
subnetwork, which can explore topological semantics from complementary
perspectives. In particular, our edge-oriented graph subnetwork utilizes the
message passing mechanism to learn neighborhood information, while our
path-oriented graph subnetwork explores high-order relationships from
substructures. To jointly learn from two subnetworks, we roughly select
informative candidate nodes with the consideration of consistency across two
subnetworks. Then, we aggregate local semantics from its K-hop subgraph based
on node degrees for topological uncertainty estimation. To overcome potential
distribution shifts, we compare target nodes and their corresponding source
nodes for discrepancy scores as an additional component for fine selection.
Extensive experiments on benchmark datasets demonstrate that DELTA outperforms
various state-of-the-art approaches.