Importance Pooling

التجميع بالأهمية

أسلوب تجميع يُرجِّح سمات الجيران بأوزان مشتقة من المشي العشوائي بدلاً من معاملتهم بالتساوي. الجيران الأكثر زيارةً يُسهمون أكثر في التمثيل المُجمَّع.

An aggregation method that weights neighbor features by random-walk-derived importance scores instead of treating all neighbors equally. More frequently visited neighbors contribute more to the pooled representation.

Also translated asالتجميع المرجَّح بالأهمية، التجميع الموزون

First appears in this corpus in: PinSage: Graph Convolutional Neural Networks for Web-Scale Recommender Systems (2018)

Appears in these papers

PinSage: Graph Convolutional Neural Networks for Web-Scale Recommender Systems2018 in the sky ✦
PinSage: Graph Convolutional Neural Networks for Web-Scale Recommender Systems2018 in the sky ✦