On Addressing the Limitations of Graph Convolutional Networks, Sitao Luan
Автор: GERAD Recherche
Загружено: 2022-11-11
Просмотров: 325
Описание:
DS4DM Coffee Talk
On Addressing the Limitations of Graph Convolutional Networks
Sitao Luan, Mcgill University
July 7, 2022
Many real-world tasks can be modeled as graphs. Recently, graph convolutional neural network (GCN) based approaches have achieved significant progress for solving large, complex, graph-structured problems. Although with high expressive power, GCNs still suffer from several difficulties, e.g. the over-smoothing problem limits deep GCNs to sufficiently exploit multi-scale information, heterophily problem makes the graph-aware models underperform the graph-agnostic models. In this presentation, I will summarize those challenges, propose some methods to address them and put forward several research problems we are currently investigating.
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке:
![Момент, когда мы перестали понимать ИИ [AlexNet]](https://ricktube.ru/thumbnail/UZDiGooFs54/mqdefault.jpg)
