Dual adaptive learning multi-task multi-view for graph network representation learning

Han, Beibei; Wei, Yingmei; Wang, Qingyong; Wan, Shanshan

doi:10.1016/j.neunet.2023.02.026

Cited by 9 publications

(9 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We compare our model with 5 representative link prediction methods: Common Neighbors (CN), Adamic Adar (AA), Node2Vec [Grover and Leskovec, 2016], GCN [Kipf and Welling, 2017], SEAL [Zhang et al, 2021]. Among them, CN and AA are heuristic Methods, and Node2Vec is a representative graph embedding method.…”

Section: Experimental Settingsmentioning

confidence: 99%

“…We leverage Pytorch Geometric to implement NCPNet and other GNN methods including GCN [Kipf and Welling, 2017] and MLP. We use the code of SEAL from the referenced paper [Zhang et al, 2021]. For graph embedding methods and Node2Vec [Grover and Leskovec, 2016], we obatain node embeddings by GraphVite 4 and use dot product to calculate connection similarity.…”

Section: Experimental Settingsmentioning

confidence: 99%

See 1 more Smart Citation

Inferring Neuron-level Brain Circuit Connection via Graph Neural Network Amidst Small Established Connections

Wan

Liao

Dong

et al. 2023

Preprint

View full text Add to dashboard Cite

Reconstructing neuron-level brain circuit network is a universally recognized formidable task. A significant impediment involves discerning the intricate interconnections among multitudinous neurons in a complex brain network. However, the majority of current methodologies only rely on learning local visual synapse features while neglecting the incorporation of comprehensive global topological connectivity information. In this paper, we consider the perspective of network connectivity and introduce graph neural networks to learn the topological features of brain networks. As a result, we propose Neuronal Circuit Prediction Network (NCPNet), a simple and effective model to jointly learn node structural representation and neighborhood representation, constructing neuronal connection pair feature for inferring neuron-level connections in a brain circuit network. We use a small number of connections randomly selected from a single brain circuit network as training data, expecting NCPNet to extrapolate known connections to unseen instances. We evaluated our model on \textit{Drosophila} connectome and \textit{C. elegans} worm connectome. The numerical results demonstrate that our model achieves a prediction accuracy of 91.88\% for neuronal connections in the \textit{Drosophila} connectome when utilizing only 5\% of known connections. Similarly, under the condition of 5\% known connections in \textit{C. elegans}, our model achieves an accuracy of 93.79\%. Additional qualitative analysis conducted on the learned representation vectors of Kenyon cells indicates that NCPNet successfully acquires meaningful features that enable the discrimination of neuronal sub-types. Our project is available at \url{https://github.com/mxz12119/NCPNet}

show abstract

Section: Experimental Settingsmentioning

confidence: 99%

Section: Experimental Settingsmentioning

confidence: 99%

Inferring Neuron-level Brain Circuit Connection via Graph Neural Network Amidst Small Established Connections

Wan

Liao

Dong

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…A similar framework, RED-GNN (Zhang and Yao, 2022) also attempts to take advantage of dynamic programming to aggregate path information. Both A * Net (Zhu et al, 2022) and AdaProp (Zhang et al, 2023) attempt to prove upon the efficiency of the previous methods by learning which nodes to propagate to.…”

Section: Effect Of Degree Messagesmentioning

confidence: 99%

“…The performance and efficiency when combining both methods is detailed in Section 4.1 and 4.2, respectively. Lastly, we note that we don't consider combining with the pruning strategy in AdaProp (Zhang et al, 2023) due to its strong similarity with that of A * Net.…”

Section: B3 Tagnet + a * Netmentioning

confidence: 99%

“…Furthermore, it inhibits their ability to propagate deeply in the graph. Two recent methods have been proposed to address the inefficiency problem, i.e., A * Net (Zhu et al, 2022) and AdaProp (Zhang et al, 2023), by only propagating to a subset of nodes every iteration. However, they still tend to propagate unnecessary and redundant messages.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Distance-Based Propagation for Efficient Knowledge Graph Reasoning

Shomer,

Ma,

et al. 2023

Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

Knowledge graph completion (KGC) aims to predict unseen edges in knowledge graphs (KGs), resulting in the discovery of new facts. A new class of methods have been proposed to tackle this problem by aggregating path information. These methods have shown tremendous ability in the task of KGC. However they are plagued by efficiency issues. Though there are a few recent attempts to address this through learnable path pruning, they often sacrifice the performance to gain efficiency. In this work, we identify two intrinsic limitations of these methods that affect the efficiency and representation quality. To address the limitations, we introduce a new method, TAGNet, which is able to efficiently propagate information. This is achieved by only aggregating paths in a fixed window for each source-target pair. We demonstrate that the complexity of TAGNet is independent of the number of layers. Extensive experiments demonstrate that TAGNet can cut down on the number of propagated messages by as much as 90% while achieving competitive performance on multiple KG datasets 1 .

show abstract

IMPRL-Net: interpretable multi-view proximity representation learning network

Lan,

Fang,

et al. 2024

Neural Comput & Applic

View full text Add to dashboard Cite

Dual adaptive learning multi-task multi-view for graph network representation learning

Cited by 9 publications

References 54 publications

Inferring Neuron-level Brain Circuit Connection via Graph Neural Network Amidst Small Established Connections

Inferring Neuron-level Brain Circuit Connection via Graph Neural Network Amidst Small Established Connections

Distance-Based Propagation for Efficient Knowledge Graph Reasoning

IMPRL-Net: interpretable multi-view proximity representation learning network

Contact Info

Product

Resources

About