基于深度学习的会话推荐方法综述
Review of Deep Learning-based Conversation Recommendation Methods
袁凤源 1梅红岩 1温民伟 1白杨 1吴帅甫1
作者信息
- 1. 辽宁工业大学 电子与信息工程学院,辽宁 锦州 121001
- 折叠
摘要
随着信息技术和智能应用的迅速发展,现今社会面临着大量数据和信息过载的挑战.为解决这一难题,基于会话的推荐方法应运而生.基于深度学习的会话推荐方法利用其强大的表示学习能力,更准确地预测用户的短期兴趣,并提供个性化的推荐服务.故综述了深度学习会话推荐方法的研究进展,包括基于卷积神经网络、图神经网络、注意力机制、多层感知器、混合模型等方法.总结了研究难点和未来研究方向.随着深度学习技术的不断进步,会话推荐方法将在实际应用中发挥越来越重要的作用.
Abstract
As information technology and intelligent applications continues to advance quickly,today's society is facing challenges such as data and information overload.To solve this problem,conversation-based recommendation methods have emerged.Based on deep learning,these methods use their powerful representation learning ability to more accurately predict users'short-term interests and provide personalized recommendation services.This article reviews the research progress of deep learning-based conversation recommendation methods,including methods based on convolutional neural networks,graph neural networks,attention mechanisms,multi-layer perceptrons,and hybrid models.The article summarizes the research challenges and future directions.As deep learning technology continues to make progress,conversation-based recommendation methods are expected to become increasingly important in practical applications.
关键词
会话/推荐系统/深度学习/机器学习/神经网络Key words
conversation/recommendation systems/deep learning/machine learning/neural network引用本文复制引用
出版年
2024
NETL
NSTL
维普
万方数据