计算机科学 ›› 2023, Vol. 50 ›› Issue (11A): 220800179-5.doi: 10.11896/jsjkx.220800179
周明星, 闫湘洲, 于敬, 高昌举, 陈运文, 纪达麒, 金克
ZHOU Mingxing, YAN Xiangzhou, YU Jing, GAO Changju, CHEN Yunwen, JI Daqi, JIN Ke
摘要： 搜索提示自动补全是正式提交搜索之前,影响用户输入搜索内容的关键手段之一,是商业搜索引擎不可或缺的核心功能之一。如何提供更好的提示词,是一个排序问题。在机器学习排序领域,收集的训练数据有位置偏差,且会影响训练模型的排序效果,已经是一个较为普遍的认知。针对以上训练数据有偏问题,对位置偏差和相关度使用深度学习分别建模,并结合改进后的上下文语义特征,新设计一种同时学习位置偏差和提示词相关度的深度学习排序算法(An Unbiased Deep Learning To Rank Algorithm for Suggestion Auto-completion,UDLTR-SAc)提升搜索提示自动补全的排序效果。UDLTR-SAc能自动学习训练数据中由于位置引入的偏差,从而学习到更为准确的相关度计算模型,在与没有考虑有偏问题的同类型算法及经典补全排序算法对比上分别获得显著增长;同时,在线上A/B测试上也获得+0.1%(p<0.1)的GMV增长。
中图分类号:
| [1]AI Q,BI K,LUO C,et al.Unbiased learning to rank with unbia-sed propensity estimation[C]//The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval.2018:385-394. [2]WANG X,BENDERSKY M,METZLER D,et al.Learning to rank with selection bias in personal search[C]//Proceedings of the 39th International ACM SIGIR Conference on Research and Development in Information Retrieval.2016:115-124. [3]WANG X,GOLBANDI N,BENDERSKY M,et al.Position bias estimation for unbiased learning to rank in personal search[C]//Proceedings of the Eleventh ACM International Conference on Web Search and Data Mining.2018:610-618. [4]HU Z,WANG Y,PENG Q,et al.Unbiased lambdamart:an unbiased pairwise learning-to-rank algorithm[C]//The World Wide Web Conference.2019:2830-2836. [5]YUE Y,JOACHIMS T.Interactively optimizing information retrieval systems as a dueling bandits problem[C]//Proceedings of the 26th Annual International Conference on Machine Lear-ning.2009:1201-1208. [6]SCHUTH A,OOSTERHUIS H,WHITESON S,et al.Mul-tileave gradient descent for fast online learning to rank[C]//Proceedings of the Ninth ACM International Conference on Web Search and Data Mining.2016:457-466. [7]WANG H,LANGLEY R,KIM S,et al.Efficient exploration of gradient space for online learning to rank[C]//The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval.2018:145-154. [8]OOSTERHUIS H,DE RIJKE M.Differentiable unbiased online learning to rank[C]//Proceedings of the 27th ACM InternationalConference on Information and Knowledge Management.2018:1293-1302. [9]YUAN K,KUANG D.Deep Pairwise Learning To Rank For Search Autocomplete[J].arXiv:2108.04976,2021. [10]BAR-YOSSEF Z,KRAUS N.Context-sensitive query auto-completion[C]//Proceedings of the 20th International Conference on World Wide Web.2011:107-116. [11]CHEN T,GUESTRIN C.Xgboost:A scalable tree boosting system[C]//Proceedings of the 22nd ACM Sigkdd International Conference on Knowledge Discovery and Data Mining.2016:785-794. [12]LI Z Y,YANG W,XIE Z J.A survey of PageRank research[J].Computer Science,2011,38(B10):185-188. [13]WANG X P,LI Z Z.An improved ranking algorithmfor web page searching[J].Computer Science,2004,31(9A). [14]LAI X X,HAN L X,ZENG X Q,et al.Research on meta-search ranking algorithms based on information quantity and information entropy[J].Computer Science,2012,39(3):153-156. [15]WEN K M,LU Z D,SUN X L,et al.A survey of semanticsearch research[J].Computer Science,2008,35(5):1-4. [16]ZHANG X,QU Y Z.Ranking problems in the Semantic Web[J].Computer Science,2008,35(2):5. [17]CARPINETO C,ROMANO G.A survey of automatic query expansion in information retrieval[J].Acm Computing Surveys(CSUR),2012,44(1):1-50. [18]LIU T Y.Learning to rank for information retrieval[J].Foundations and Trends© in Information Retrieval,2009,3(3):225-331. [19]CAI F,DE RIJKE M.A survey of query auto completion in information retrieval[J].Foundations and Trends© in Information Retrieval,2016,10(4):273-363. [20]JOACHIMS T,SWAMINATHAN A,SCHNABEL T.Unbi-ased learning-to-rank with biased feedback[C]//Proceedings of the Tenth ACM International Conference on Web Search and Data Mining.2017:781-789. [21]GUO H,YU J,LIU Q,et al.PAL:a position-bias aware learning framework for CTR prediction in live recommender systems[C]//Proceedings of the 13th ACM Conference on Recommender Systems.2019:452-456. |
|
||
首页