Friend Recommendations Retrieval From Graph Search to Embedding based
Автор: RecSys ML
Загружено: 2024-11-24
Просмотров: 1437
We discuss ideas of the state of the art in friend recommendation retrieval in social networks by leaders of the industry. These ideas have led to 10+ percentage growth in connections in multiple published papers and can be highly additive to the success of your social network applications.
0:00 what is Friend recommendations
1:00 friends of friends as a way to recommend new friends
4:07 what is success for retrieval
7:35 Recap so far and weighted paths in graph search
11:58 Deepwalk as a way to generate embeddings based on sampling walks
14:00 Clustering as a way to view social graph and recommend friends and also run experiments without network interference bias
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
