大型语料库项目推荐的抽样偏差校正神经模型

Xinyang Yi, Ji Yang, Lichan Hong, D. Cheng, L. Heldt, A. Kumthekar, Zhe Zhao, Li Wei, Ed H. Chi
{"title":"大型语料库项目推荐的抽样偏差校正神经模型","authors":"Xinyang Yi, Ji Yang, Lichan Hong, D. Cheng, L. Heldt, A. Kumthekar, Zhe Zhao, Li Wei, Ed H. Chi","doi":"10.1145/3298689.3346996","DOIUrl":null,"url":null,"abstract":"Many recommendation systems retrieve and score items from a very large corpus. A common recipe to handle data sparsity and power-law item distribution is to learn item representations from its content features. Apart from many content-aware systems based on matrix factorization, we consider a modeling framework using two-tower neural net, with one of the towers (item tower) encoding a wide variety of item content features. A general recipe of training such two-tower models is to optimize loss functions calculated from in-batch negatives, which are items sampled from a random mini-batch. However, in-batch loss is subject to sampling biases, potentially hurting model performance, particularly in the case of highly skewed distribution. In this paper, we present a novel algorithm for estimating item frequency from streaming data. Through theoretical analysis and simulation, we show that the proposed algorithm can work without requiring fixed item vocabulary, and is capable of producing unbiased estimation and being adaptive to item distribution change. We then apply the sampling-bias-corrected modeling approach to build a large scale neural retrieval system for YouTube recommendations. The system is deployed to retrieve personalized suggestions from a corpus with tens of millions of videos. We demonstrate the effectiveness of sampling-bias correction through offline experiments on two real-world datasets. We also conduct live A/B testings to show that the neural retrieval system leads to improved recommendation quality for YouTube.","PeriodicalId":215384,"journal":{"name":"Proceedings of the 13th ACM Conference on Recommender Systems","volume":"36 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"148","resultStr":"{\"title\":\"Sampling-bias-corrected neural modeling for large corpus item recommendations\",\"authors\":\"Xinyang Yi, Ji Yang, Lichan Hong, D. Cheng, L. Heldt, A. Kumthekar, Zhe Zhao, Li Wei, Ed H. Chi\",\"doi\":\"10.1145/3298689.3346996\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Many recommendation systems retrieve and score items from a very large corpus. A common recipe to handle data sparsity and power-law item distribution is to learn item representations from its content features. Apart from many content-aware systems based on matrix factorization, we consider a modeling framework using two-tower neural net, with one of the towers (item tower) encoding a wide variety of item content features. A general recipe of training such two-tower models is to optimize loss functions calculated from in-batch negatives, which are items sampled from a random mini-batch. However, in-batch loss is subject to sampling biases, potentially hurting model performance, particularly in the case of highly skewed distribution. In this paper, we present a novel algorithm for estimating item frequency from streaming data. Through theoretical analysis and simulation, we show that the proposed algorithm can work without requiring fixed item vocabulary, and is capable of producing unbiased estimation and being adaptive to item distribution change. We then apply the sampling-bias-corrected modeling approach to build a large scale neural retrieval system for YouTube recommendations. The system is deployed to retrieve personalized suggestions from a corpus with tens of millions of videos. We demonstrate the effectiveness of sampling-bias correction through offline experiments on two real-world datasets. We also conduct live A/B testings to show that the neural retrieval system leads to improved recommendation quality for YouTube.\",\"PeriodicalId\":215384,\"journal\":{\"name\":\"Proceedings of the 13th ACM Conference on Recommender Systems\",\"volume\":\"36 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2019-09-10\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"148\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Proceedings of the 13th ACM Conference on Recommender Systems\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/3298689.3346996\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 13th ACM Conference on Recommender Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3298689.3346996","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 148

摘要

许多推荐系统从一个非常大的语料库中检索和评分项目。处理数据稀疏性和幂律项分布的常用方法是从内容特征中学习项表示。除了许多基于矩阵分解的内容感知系统外,我们还考虑了一个使用双塔神经网络的建模框架,其中一个塔(项目塔)编码各种各样的项目内容特征。训练这种双塔模型的一般方法是优化从批内负(从随机小批中抽样的项目)计算的损失函数。然而,批内损失受到抽样偏差的影响,可能会损害模型性能,特别是在高度倾斜分布的情况下。本文提出了一种从流数据中估计项目频率的新算法。理论分析和仿真结果表明,该算法不需要固定的项目词汇表,能够产生无偏估计,并能适应项目分布的变化。然后,我们应用采样偏差校正建模方法来构建一个大规模的YouTube推荐神经检索系统。该系统用于从包含数千万个视频的语料库中检索个性化建议。我们通过两个真实数据集的离线实验证明了抽样偏差校正的有效性。我们还进行了实时A/B测试,以证明神经检索系统可以提高YouTube的推荐质量。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Sampling-bias-corrected neural modeling for large corpus item recommendations
Many recommendation systems retrieve and score items from a very large corpus. A common recipe to handle data sparsity and power-law item distribution is to learn item representations from its content features. Apart from many content-aware systems based on matrix factorization, we consider a modeling framework using two-tower neural net, with one of the towers (item tower) encoding a wide variety of item content features. A general recipe of training such two-tower models is to optimize loss functions calculated from in-batch negatives, which are items sampled from a random mini-batch. However, in-batch loss is subject to sampling biases, potentially hurting model performance, particularly in the case of highly skewed distribution. In this paper, we present a novel algorithm for estimating item frequency from streaming data. Through theoretical analysis and simulation, we show that the proposed algorithm can work without requiring fixed item vocabulary, and is capable of producing unbiased estimation and being adaptive to item distribution change. We then apply the sampling-bias-corrected modeling approach to build a large scale neural retrieval system for YouTube recommendations. The system is deployed to retrieve personalized suggestions from a corpus with tens of millions of videos. We demonstrate the effectiveness of sampling-bias correction through offline experiments on two real-world datasets. We also conduct live A/B testings to show that the neural retrieval system leads to improved recommendation quality for YouTube.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信