GSM: inductive learning on dynamic graph embeddings

M Ananyeva, I Makarov, M Pendiukhov - Network Algorithms, Data Mining …, 2020 - Springer
M Ananyeva, I Makarov, M Pendiukhov
Network Algorithms, Data Mining, and Applications: NET, Moscow, Russia, May 2018 8, 2020Springer
In this paper, we study the problem of learning graph embeddings for dynamic networks and
the ability to generalize to unseen nodes called inductive learning. Firstly, we overview the
state-of-the-art methods and techniques for constructing graph embeddings and learning
algorithms for both transductive and inductive approaches. Secondly, we propose an
improved GSM based on GraphSAGE algorithm and set up the experiments on datasets
CORA, Reddit, and HSEcite, which is collected from Scopus citation database across the …
Abstract
In this paper, we study the problem of learning graph embeddings for dynamic networks and the ability to generalize to unseen nodes called inductive learning. Firstly, we overview the state-of-the-art methods and techniques for constructing graph embeddings and learning algorithms for both transductive and inductive approaches. Secondly, we propose an improved GSM based on GraphSAGE algorithm and set up the experiments on datasets CORA, Reddit, and HSEcite, which is collected from Scopus citation database across the authors with affiliation to NRU HSE in 2011–2017. The results show that our three-layer model with attention-based aggregation function, added normalization layers, regularization (dropout) outperforms suggested by the respective authors’ GraphSAGE models with mean, LSTM, and pool aggregation functions, thus giving more insight into possible ways to improve inducting learning model based on GraphSAGE model.
Springer
以上显示的是最相近的搜索结果。 查看全部搜索结果