Скачать или смотреть Learn low-dim Embeddings that encode GRAPH structure (data) : "Representation Learning" /arXiv

Optimize your complex Graph Data before applying Neural Network predictions. Automatically learn to encode graph structure into low-dimensional embeddings, using techniques based on deep learning and nonlinear dimensionality reduction.

An encoder-decoder perspective, random walk approaches or Neighborhood aggregation methods/encoders.

Since central problem in machine learning on graphs is finding a way to incorporate information about graph-structure into a machine learning model.

Find a way to represent, or encode, graph structure so that it can be easily exploited by your Machine Learning models.

Great publication:
-----------------------------
"Representation Learning on Graphs: Methods and Applications"by William L. Hamilton, Rex Ying and Jure Leskovec
arXiv:1709.05584v3 [cs.SI] 10 Apr 2018

#GraphNN
#RepresentationLearning
#EncodeGraph

00:00 Representation Learning
04:22 embed in low-dimensional vector space
06:23 Embedded nodes
08:34 Encoder-Decoder
12:15 Random walk approach
16:06 Drawbacks
17:02 Neighborhood Autoencoder
21:20 Aggregate Node's local Neighborhood
22:16 Summary