知识图谱驱动的可解释推荐规则联合学习

需积分: 9 135 浏览量更新于2024-09-08 收藏 1.34MB PDF 举报

“联合学习可解释规则的推荐系统与知识图谱” 在推荐系统领域，解释性和有效性是构建高质量系统的关键要素。以往的工作主要集中在利用侧信息来提升推荐性能。然而，这些方法存在一些不足：（1）基于神经网络的嵌入方法的预测难以解释和调试；（2）符号化的图基方法（如元路径模型）需要手动定义模式和规则，并且忽略了物品关联的类型（如替代关系和互补关系）。针对这些问题，论文提出了一个新颖的联合学习框架，将知识图谱中可解释规则的归纳与引导规则的神经推荐模型的构建相结合。该框架鼓励两个模块相互补充，共同生成有效且可解释的推荐结果：（1）归纳规则模块从以物品为中心的知识图谱中挖掘出的规则，总结了多跳关系模式，用于推断不同物品关联，并提供易于理解的模型预测解释；（2）推荐模块可以借助归纳出的规则进行增强，从而在处理冷启动问题时具有更好的泛化能力。大量的实验表明，提出的这种方法在实际世界数据集上的物品推荐任务上相对于基线方法有显著的改进。我们的模型在包含“噪声”的物品知识图谱上展现出稳健的性能，这些噪声图谱是通过链接物品名称到相关实体生成的。该研究的贡献包括： 1. 设计了一个联合学习框架，将知识图谱的规则学习与推荐模型的构建集成，解决了传统方法的解释性和泛化性问题。 2. 提出了归纳规则模块，自动发现知识图谱中的模式，减少了对人工定义规则的依赖，同时考虑了物品关联的多样性。 3. 证明了推荐模块结合归纳规则后，对于处理新物品或稀疏数据（冷启动问题）的推荐性能提升。 4. 在多个真实数据集上进行了实验验证，展示了方法的有效性和鲁棒性。这篇论文提出的方法为推荐系统领域带来了重要的进步，它通过结合神经网络的灵活性和知识图谱的结构信息，提高了推荐的可解释性和准确性。这为未来开发更加智能和用户友好的推荐系统提供了新的研究方向。

Learning Rules Rule Selection

iPhone MacBook

item

iPhone’s

charger

Apple

Entity

Apple’s

charger

MacBook

iPhone

Knowledge graph

Google

Linking items

Pixel

Relations

Buy Together

Phones.

Manufacturer

Accessories.

Manufacturer

Accessories.

Manufacturer

Also View

Phones.

Manufacturer

Rivals

Buy Together

Rules 𝒘

(

phones.manufacturer,

accessories.manufactur

e𝑟

)

→&Buy Together

0.12

(

phones.manufacturer, rivals,

phones.manufacture

𝑟

)

→ Also View

0.21

(

phones.manufacturer, rivals,

laptaops.manufactur

𝑒𝑟

)

→&Also View

0.03

(

phones.manufacturer,

accessories.

earsets..manufacture𝑟

)

→&Buy Also

0.14

…

Purchase

history

iPhone

Battery

Monitor

…

Rule Learning Module

Rules

Recommended

items

Earset

Laptop

….

Recommendation Module

Black edges: Relations

Colored edges: Item associations

𝒘

0.12

0.21

0.03

0.14

…

Heterogeneous Graph Construction

Figure 2: Overview of the Proposed RuleRec Framework

. First, we build a heterogeneous graph from items and a knowledge graph. The rule learning

module learns the importance of rules and the recommendation module learns the importance at the same time by sharing a parameter vector w.

Entities

Items

(

)

Paths between ! and ":

, -

(

)

Edge type:

6 -

(

6 -

7 8

6 -

7 8

6 -

7 8

(

Related reasoning rules:

Rules for *

9: ;

, 8

Rules for *

6 *

9: ;

, 8

Figure 3:

An example of a heterogeneous graph which consists of items

and entities in a knowledge graph. The dashed lines are links between items

and entities generated by an entity linking algorithm.

respectively. The prediction equation of NCF is dened in Eq

(5)

in which the outputs of GMF and MLP parts are concatentated to

get the nal score. And we modied the objective function of NCF

into Eq (6) in this paper.

u, i

= ϕ(α · h

u, i

⊕ (1 − α) · g

u, i

) (5)

N C F

= σ (

u ∈U

p ∈I

,n <I

u, p

− S

u, n

))

(6)

3 THE RULEREC FRAMEWORK

Framework Overview.

Recommendation with rule learning con-

sists of two sub-tasks: 1) rule learning in a knowledge graph based

on item associations; 2) recommending items for each user

with

his/her purchase history I

and the derived rules R.

To cope with these tasks, we design a multi-task learning frame-

work. The framework consists of two modules, a rule learning

module and a recommendation module. The rule learning module

aims to derive useful rules through reasoning rules with ground-

truth item associations in the knowledge graph. Based on the rule

set, we can generate an item-pair feature vector whose each entry

is an encoded value of each rule. The recommendation module

takes the item-pair feature vector as additional input to enhance

recommendation performances and give explanations for the rec-

ommendation. We introduce a shared rule weight vector

which

indicates the importance of each rule in predicting user preference,

and shows the eectiveness of each rule in predicting item pair

associations. Besides, based on the assume that useful rules per-

form consistently in both modules with higher weights, we design

a objective function to conduct jointly learning:

min

V ,W

O = min

V ,W

+ λO

} (7)

where

denotes the parameters of the recommendation module,

and

represents the shared parameters of the rule learning and

the recommendation module. The objective function consists of two

terms:

is the objective of the recommendation module, which

recommends items based on the induced rules.

is the objective

of the rule learning module, in which we leverage the given item

associations to learn useful rules. λ is a trade-o parameter.

3.1 Heterogeneous Graph Construction

First, we build a heterogeneous graph containing items for the

recommendation and a knowledge graph. For some items, we can

conduct exactly mapping between the item and the entity, such as

“iPhone", “Macbook". For other items, it is hard to nd an entity

that represents the items, such iPhone’s charger. Thus, we adopt

entity linking algorithm [

] to nd the related entities of an item

from its title, brand, and description in the shopping website. In this

way, we can add new nodes to the knowledge graph that represents

items and add some edges for it according to entity linking results.

Then, we get a heterogeneous graph which contains the items and

the original knowledge graph. Fig. 3 is an example.

3.2 Rule Learning Module

The rule learning module aims to nd the reliable rule set

asso-

ciated with given item associations A in the heterogeneous graph.

Rule learning.

For any item pair (

) in the heterogeneous graph,

we use a random walk based algorithm to compute the probabilities

of nding paths which follow certain rules between the item pair,

similar to [

]. Then, we obtain feature vectors for item pairs.

Each entry of the feature vector is the probability of a rule between

the item pair. Here, we focus on relation types between the item

pair to obtain rules such as

in Fig. 3, because it is general to the

entities to capture the rules between items.

剩余10页未读，继续阅读

Jayxp

粉丝: 6
资源: 137

知识图谱驱动的可解释推荐规则联合学习

Label Efficient Semi-Supervised Learning via Graph Filtering.pdf

Neural Machine Translation by Jointly Learning to Align and Translate.pdf

Jointly Learning Binary Code for Large-scale Face Image Retrieval and Attributes

Classification with color and__ texture jointly or separately AA.pdf

Joint Multilingual Supervision for Cross-lingual Entity Linking.pdf

Neural machine translation by jointly learning to align and translate

Neural Machine Translation by Jointly Learning to Align and Translate阅读笔记

13、1604.01655.pdf

SuperGlue: Learning Feature Matching with Graph Neural Networks特

供应链管理.pdf

最新资源