HKUDS / MAERec

[SIGIR'2023] "MAERec: Graph Masked Autoencoder for Sequential Recommendation"
https://arxiv.org/abs/2305.04619
55 stars 5 forks source link

Questions about algorithmic thinking #5

Closed Keplerrrrrrrr closed 1 year ago

Keplerrrrrrrr commented 1 year ago

Hello, thanks for your outstanding work! I am interested in your paper "Adaptive Path Masking", the idea of expanding and sampling the nodes in the graph. Is it a kind of Heuristically Search?

HelloElwin commented 1 year ago

Hi, thanks for the question! I think there's a difference between them. Expanding and sampling are used to generate the transition paths from anchor nodes. The process stops after a few steps (as specified in the hyper-parameters) and it is not 'searching' for something. Instead, it simply produces randomly generated transition paths from anchor nodes.