Graph Learning based Recommender Systems: A ReviewyIJCAI2021一篇基于图学习的推荐系统综述

原文链接Graph Learning based Recommender Systems: A Review (ijcai.org)

 根据RS中的数据分类

GLRS Built on General Interaction Data

 挑战：it is challenging how to effectively and effificiently propagate the information between users or items .This is particularly challenging in a bipartite graph since no direct links exist between users or items, and thus the information should be propagated via multi-hop neighbour nodes.

GLRS Built on Sequential Interaction Data 

交互序列表示： For a given user u, a single-type interaction sequence is usually recorded as a sequence of interacted with items (denoted as v ), e.g., {v1, ..., vn} , while a multi-type interaction sequence is recorded as a sequence of <interaction type, item > pairs, e.g., {click v1, click v2, ..., purchase vn} .

序列化交互数据表示为一个有向图，图中每条路径代表一个交互序列。在每条路径中，交互作为节点，相邻节点之间的有向边表示交互的顺序。 一个用户可能在一个序列中发生多个相同的交互（例如，多次单击相同的项目），从而导致由一个或多个循环组成的路径。

优点：the strong capability of graph learning to represent and model even the most complicated transitions in a sequence of interactions. 

挑战：In particular it is critical how to construct a graph to effectively represent the sequential interaction data with minimal information loss, and how to propagate information on the graph to effectively model even the most complicated transitions.

 

GLRS Incorporating Side Information Data 

(1) attribute information

三种结点类型： user node、 item node、attribute value node

边：user-item edges 、user (or item)-attribute value edge

挑战：it is challenging to selectively aggregate those useful attribute information to improve the recommendation performance.

(2) social information

两部分构成的异质图：u-i二部图+社交网络图

The combination of social information and general or sequential user-item interaction data naturally results in a heterogeneous graph comprising two parts. The first is the bipartite graph derived from the general interaction data  or the directed graph extracted from the sequential interaction data, while the second part is the social graph connecting the users.

挑战：on one hand, it is not clear how many orders of neighbours should be considered to correctly compute this inflfluence on a given user. On the other hand, different neighbours usually inflfluence a user to different degrees . Hence, it is a challenge to appropriately model the inflfluence of other users to a given user.

 另外一个应用社交网络推荐Friend recommendation：challenge lies in how to appropriately model the mutual inflfluence between users

(3) external knowledge

① incorporating ontology knowledge.

it remains a challenge to propagate users’ preferences over items along the hierarchy tree graph to extract the multi-level preferences.

②incorporating common knowledge.

 However, it remains a challenge to effectively propagate information between different types of entities via different types of links between them.