推荐中的序列化建模:Session-based neural recommendation

2017 年 11 月 5 日 机器学习研究会 白婷
推荐中的序列化建模:Session-based neural recommendation


点击上方 “机器学习研究会”可以订阅
摘要
 

转自:白婷

首先介绍下session-based 的概念:session是服务器端用来记录识别用户的一种机制. 典型的场景比如购物车,服务端为特定的对象创建了特定的Session,用于标识这个对象,并且跟踪用户的浏览点击行为。我们这里可以将其理解为具有时序关系的一些记录序列。


一、写作动机

传统的两类推荐方法——基于内容的推荐算法和协同过滤推荐算法(model-based、memory-based)在刻画序列数据中存在缺陷:每个item相互独立,不能建模session中item的连续偏好信息。


二、传统的解决方法

  1. item-to-item recommendation approach (Sarwar et al.,2001; Linden et al., 2003) : 采用session中item间的相似性预测下一个item。缺点:只考虑了最后一次的click 的item相似性, 忽视了前面的的clicks, 没有考虑整个序列信息。

  2.  Markov decision Processes (MDPs)(Shani et al., 2002):马尔科夫决策过程,用四元组<S,A, P, R>(S: 状态, A: 动作, P: 转移概率, R: 奖励函数)刻画序列信息,通过状态转移概率的计算点击下一个动作:即点击item的概率。缺点:状态的数量巨大,会随问题维度指数增加。(MDPs参见博客:增强学习(二)----- 马尔可夫决策过程MDP


三、Deep Neural Network的方法

Deep Neural Network (RNN:LSTM和GRU的记忆性) 被成功的应用在刻画序列信息。因为论文中主要采用GRU,下面简单介绍下GRU.(LSTM 详解参加博客:Understanding LSTM Networks)


原文链接:

https://zhuanlan.zhihu.com/p/30720579?group_id=910073604973465600

“完整内容”请点击【阅读原文】
↓↓↓
登录查看更多
14

相关内容

序列化 (Serialization)将对象的状态信息转换为可以存储或传输的形式的过程。

Model-based methods for recommender systems have been studied extensively in recent years. In systems with large corpus, however, the calculation cost for the learnt model to predict all user-item preferences is tremendous, which makes full corpus retrieval extremely difficult. To overcome the calculation barriers, models such as matrix factorization resort to inner product form (i.e., model user-item preference as the inner product of user, item latent factors) and indexes to facilitate efficient approximate k-nearest neighbor searches. However, it still remains challenging to incorporate more expressive interaction forms between user and item features, e.g., interactions through deep neural networks, because of the calculation cost. In this paper, we focus on the problem of introducing arbitrary advanced models to recommender systems with large corpus. We propose a novel tree-based method which can provide logarithmic complexity w.r.t. corpus size even with more expressive models such as deep neural networks. Our main idea is to predict user interests from coarse to fine by traversing tree nodes in a top-down fashion and making decisions for each user-node pair. We also show that the tree structure can be jointly learnt towards better compatibility with users' interest distribution and hence facilitate both training and prediction. Experimental evaluations with two large-scale real-world datasets show that the proposed method significantly outperforms traditional methods. Online A/B test results in Taobao display advertising platform also demonstrate the effectiveness of the proposed method in production environments.

0
8
下载
预览

Given e-commerce scenarios that user profiles are invisible, session-based recommendation is proposed to generate recommendation results from short sessions. Previous work only considers the user's sequential behavior in the current session, whereas the user's main purpose in the current session is not emphasized. In this paper, we propose a novel neural networks framework, i.e., Neural Attentive Recommendation Machine (NARM), to tackle this problem. Specifically, we explore a hybrid encoder with an attention mechanism to model the user's sequential behavior and capture the user's main purpose in the current session, which are combined as a unified session representation later. We then compute the recommendation scores for each candidate item with a bi-linear matching scheme based on this unified session representation. We train NARM by jointly learning the item and session representations as well as their matchings. We carried out extensive experiments on two benchmark datasets. Our experimental results show that NARM outperforms state-of-the-art baselines on both datasets. Furthermore, we also find that NARM achieves a significant improvement on long sessions, which demonstrates its advantages in modeling the user's sequential behavior and main purpose simultaneously.

0
5
下载
预览
小贴士
相关资讯
LibRec 精选:基于LSTM的序列推荐实现(PyTorch)
LibRec智能推荐
44+阅读 · 2018年8月27日
【推荐】RNN最新研究进展综述
机器学习研究会
22+阅读 · 2018年1月6日
【推荐系统】融合 MF 和 RNN 的电影推荐系统
产业智能官
23+阅读 · 2018年1月4日
融合 MF 和 RNN 的电影推荐系统
PaperWeekly
7+阅读 · 2017年12月28日
赛尔原创 | 基于转移的语义依存图分析
哈工大SCIR
5+阅读 · 2017年11月20日
【推荐】RNN/LSTM时序预测
机器学习研究会
24+阅读 · 2017年9月8日
相关VIP内容
和积网络综述论文,Sum-product networks: A survey,24页pdf
专知会员服务
21+阅读 · 2020年4月3日
专知会员服务
92+阅读 · 2020年2月22日
专知会员服务
66+阅读 · 2020年1月20日
可解释推荐:综述与新视角
专知会员服务
81+阅读 · 2019年10月13日
相关论文
Qingyu Guo,Fuzhen Zhuang,Chuan Qin,Hengshu Zhu,Xing Xie,Hui Xiong,Qing He
82+阅读 · 2020年2月28日
Xiang Wang,Xiangnan He,Meng Wang,Fuli Feng,Tat-Seng Chua
8+阅读 · 2019年5月20日
Chih-Ming Chen,Chuan-Ju Wang,Ming-Feng Tsai,Yi-Hsuan Yang
6+阅读 · 2019年2月19日
Next Item Recommendation with Self-Attention
Shuai Zhang,Yi Tay,Lina Yao,Aixin Sun
5+阅读 · 2018年8月25日
Han Zhu,Xiang Li,Pengye Zhang,Guozheng Li,Jie He,Han Li,Kun Gai
8+阅读 · 2018年5月21日
Lei Zheng,Chun-Ta Lu,Lifang He,Sihong Xie,Vahid Noroozi,He Huang,Philip S. Yu
5+阅读 · 2018年5月18日
Yongfeng Zhang,Xu Chen
10+阅读 · 2018年5月13日
Tran Dang Quang Vinh,Tuan-Anh Nguyen Pham,Gao Cong,Xiao-Li Li
12+阅读 · 2018年4月18日
Qiang Cui,Shu Wu,Yan Huang,Liang Wang
5+阅读 · 2017年12月7日
Jing Li,Pengjie Ren,Zhumin Chen,Zhaochun Ren,Jun Ma
5+阅读 · 2017年11月13日
Top