Matches in SemOpenAlex for { <https://semopenalex.org/work/W3162116716> ?p ?o ?g. }
- W3162116716 abstract "Graph Neural Networks (GNNs) have emerged as a flexible and powerful approach for learning over graphs. Despite this success, existing GNNs are constrained by their local message-passing architecture and are provably limited in their expressive power. In this work, we propose a new GNN architecture -- the Neural Tree. The neural tree architecture does not perform message passing on the input graph but on a tree-structured graph, called the H-tree, that is constructed from the input graph. Nodes in the H-tree correspond to subgraphs in the input graph, and they are reorganized in a hierarchical manner such that a parent-node of a node in the H-tree always corresponds to a larger subgraph in the input graph. We show that the neural tree architecture can approximate any smooth probability distribution function over an undirected graph, as well as emulate the junction tree algorithm. We also prove that the number of parameters needed to achieve an $epsilon$-approximation of the distribution function is exponential in the treewidth of the input graph, but linear in its size. We apply the neural tree to semi-supervised node classification in 3D scene graphs, and show that these theoretical properties translate into significant gains in prediction accuracy, over the more traditional GNN architectures." @default.
- W3162116716 created "2021-05-24" @default.
- W3162116716 creator A5022339346 @default.
- W3162116716 creator A5042157108 @default.
- W3162116716 creator A5047680302 @default.
- W3162116716 creator A5076626581 @default.
- W3162116716 date "2021-05-15" @default.
- W3162116716 modified "2023-09-27" @default.
- W3162116716 title "Neural Trees for Learning on Graphs" @default.
- W3162116716 cites W1501856433 @default.
- W3162116716 cites W1511986666 @default.
- W3162116716 cites W1587990862 @default.
- W3162116716 cites W1902387477 @default.
- W3162116716 cites W1905882502 @default.
- W3162116716 cites W1948169740 @default.
- W3162116716 cites W1964821516 @default.
- W3162116716 cites W1980452149 @default.
- W3162116716 cites W1990761146 @default.
- W3162116716 cites W2070155084 @default.
- W3162116716 cites W2077069816 @default.
- W3162116716 cites W2116341502 @default.
- W3162116716 cites W2116793595 @default.
- W3162116716 cites W2150120952 @default.
- W3162116716 cites W2154837415 @default.
- W3162116716 cites W2277195237 @default.
- W3162116716 cites W2396147015 @default.
- W3162116716 cites W2479423890 @default.
- W3162116716 cites W2506483933 @default.
- W3162116716 cites W2550848904 @default.
- W3162116716 cites W2558748708 @default.
- W3162116716 cites W2579549467 @default.
- W3162116716 cites W2606780347 @default.
- W3162116716 cites W2787705367 @default.
- W3162116716 cites W2811124557 @default.
- W3162116716 cites W2886970679 @default.
- W3162116716 cites W2889399096 @default.
- W3162116716 cites W2894982307 @default.
- W3162116716 cites W2900470550 @default.
- W3162116716 cites W2906340703 @default.
- W3162116716 cites W2906413107 @default.
- W3162116716 cites W2911710347 @default.
- W3162116716 cites W2920448302 @default.
- W3162116716 cites W2962711740 @default.
- W3162116716 cites W2962767366 @default.
- W3162116716 cites W2962810718 @default.
- W3162116716 cites W2962988969 @default.
- W3162116716 cites W2963312446 @default.
- W3162116716 cites W2963536419 @default.
- W3162116716 cites W2963649796 @default.
- W3162116716 cites W2963819570 @default.
- W3162116716 cites W2963858333 @default.
- W3162116716 cites W2963907666 @default.
- W3162116716 cites W2964015378 @default.
- W3162116716 cites W2964113829 @default.
- W3162116716 cites W2964321699 @default.
- W3162116716 cites W2967353954 @default.
- W3162116716 cites W2970482593 @default.
- W3162116716 cites W2970493342 @default.
- W3162116716 cites W2971078096 @default.
- W3162116716 cites W2979824638 @default.
- W3162116716 cites W2987119394 @default.
- W3162116716 cites W2990045899 @default.
- W3162116716 cites W2990129662 @default.
- W3162116716 cites W2995889583 @default.
- W3162116716 cites W2996788836 @default.
- W3162116716 cites W3035154952 @default.
- W3162116716 cites W3035356236 @default.
- W3162116716 cites W3035664258 @default.
- W3162116716 cites W3039925688 @default.
- W3162116716 cites W3099732281 @default.
- W3162116716 cites W3110824791 @default.
- W3162116716 cites W3119755129 @default.
- W3162116716 cites W3123260146 @default.
- W3162116716 cites W3203511368 @default.
- W3162116716 cites W637153065 @default.
- W3162116716 cites W2570241400 @default.
- W3162116716 hasPublicationYear "2021" @default.
- W3162116716 type Work @default.
- W3162116716 sameAs 3162116716 @default.
- W3162116716 citedByCount "1" @default.
- W3162116716 countsByYear W31621167162021 @default.
- W3162116716 crossrefType "posted-content" @default.
- W3162116716 hasAuthorship W3162116716A5022339346 @default.
- W3162116716 hasAuthorship W3162116716A5042157108 @default.
- W3162116716 hasAuthorship W3162116716A5047680302 @default.
- W3162116716 hasAuthorship W3162116716A5076626581 @default.
- W3162116716 hasConcept C100560664 @default.
- W3162116716 hasConcept C113174947 @default.
- W3162116716 hasConcept C11413529 @default.
- W3162116716 hasConcept C114614502 @default.
- W3162116716 hasConcept C127413603 @default.
- W3162116716 hasConcept C132525143 @default.
- W3162116716 hasConcept C132569581 @default.
- W3162116716 hasConcept C134306372 @default.
- W3162116716 hasConcept C147517495 @default.
- W3162116716 hasConcept C151376022 @default.
- W3162116716 hasConcept C154945302 @default.
- W3162116716 hasConcept C163797641 @default.
- W3162116716 hasConcept C197855036 @default.
- W3162116716 hasConcept C203776342 @default.