利用对话模型引导的对话生成推荐

doi:10.12068/j.issn.1005-3026.2022.10.005

东北大学学报（自然科学版） ›› 2022, Vol. 43 ›› Issue (10): 1397-1404.DOI: 10.12068/j.issn.1005-3026.2022.10.005

利用对话模型引导的对话生成推荐

齐孝龙，韩东红，高翟，乔百友

(东北大学计算机科学与工程学院，辽宁沈阳110169)

修回日期:2021-10-11 接受日期:2021-10-11 发布日期:2022-11-07
通讯作者: 齐孝龙
作者简介:齐孝龙(1996-)，男，河南安阳人，东北大学硕士研究生；韩东红(1968-)，女，河北平山人，东北大学教授，博士生导师.
基金资助:
国家重点研发计划项目(2019YFB1405300)；国家自然科学基金资助项目(61672144)；中央高校基本科研业务费专项资金资助项目(N2016009).

Using Dialogue Model to Guide Recommendation of Dialogue Generation

QI Xiao-long， HAN Dong-hong， GAO Di， QIAO Bai-you

School of Computer Science & Engineering， Northeastern University， Shenyang 110169， China.

Revised:2021-10-11 Accepted:2021-10-11 Published:2022-11-07
Contact: HAN Dong-hong
About author:-
Supported by:
-

摘要/Abstract

摘要： 对话推荐技术旨在通过与用户的对话交互完成高质量的信息推荐.针对已有研究存在的对话目标预测准确性不高的问题，提出一种利用对话模型引导的对话生成推荐(dialogue guided recommendation of dialogue generation， DGRDG)模型.首先，利用对话模型生成对话目标，通过经典的Seq2Seq模型融合输入的对话历史、用户画像以及知识信息来生成对话目标；其次，提出目标重规划策略(goal replan policy， GRP)来修正生成的对话目标，以提高对话目标预测的准确率.在DuRecDial数据集上进行实验的结果表明，对话目标生成模块在引入目标重规划策略后，对话目标预测的准确率提高了3.93%；总体模型在BLEU，DISTINCT，F1以及人工评价指标上具有较好的效果.

关键词: 对话推荐；对话生成；对话目标规划；管道模型；对话策略

Abstract: Conversational recommendation technology aims to achieve high-quality information recommendation through dialogue interaction with users. Aiming at the problem that the accuracy of dialogue goal prediction is not high， a dialogue guided recommendation of dialogue generation(DGRDG)model is proposed. Firstly， a dialogue model is used to generate the dialogue goal， and the classic Seq2Seq model is used to fuse the input dialogue history， user profile and knowledge information to generate the dialogue goal. Secondly， a goal replan policy(GRP)is proposed to modify the generated dialogue goal to improve the accuracy of dialogue goal prediction. The experimental results on DuRecDial dataset show that the accuracy of dialogue goal prediction is improved by 3.93% after the GRP is introduced into the dialogue goal generation module. And the overall model has acquired good results in BLEU， DISTINCT， F1 and human evaluation.

Key words: conversational recommendation; dialogue generation; dialogue goal planning; pipeline model; dialogue policy

中图分类号:

TP20

齐孝龙，韩东红，高翟，乔百友. 利用对话模型引导的对话生成推荐[J]. 东北大学学报（自然科学版）, 2022, 43(10): 1397-1404.

QI Xiao-long， HAN Dong-hong， GAO Di， QIAO Bai-you. Using Dialogue Model to Guide Recommendation of Dialogue Generation[J]. Journal of Northeastern University(Natural Science), 2022, 43(10): 1397-1404.

参考文献

[1]秦川，祝恒书，庄福振，等.基于知识图谱的推荐系统研究综述［J］.中国科学:信息科学，2020，50(7):937-956.(Qin Chuan，Zhu Heng-shu，Zhuang Fu-zhen，et al A review of research on recommendation system based on knowledge map ［J］.Chinese Science:Information Science，2020，50(7):937-956.)
[2]Li S，Lei W，Wu Q，et al.Seamlessly unifying attributes and items:conversational recommendation for cold-start users［J］.ACM Transactions on Information Systems(TOIS)，2021，39(4):1-29.
[3]Chen Z，Wang X，Xie X，et al.Towards explainable conversational recommendation［C］//International Joint Conference on Artificial Intelligence(IJCAI).Yokohama，2020:2994-3000.
[4]Li R，Kahou S，Schulz H，et al.Towards deep conversational recommendations［C］//Proceedings of the 32nd International Conference on Neural Information Processing Systems.Montreal，2018:9748-9758.
[5]Chen Q，Lin J，Zhang Y，et al.Towards knowledge-based recommender dialog system［C］//Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing(EMNLP-IJCNLP).Hong Kong，2019:1803-1813.
[6]Zhou K，Zhao W X，Bian S，et al.Improving conversational recommender systems via knowledge graph based semantic fusion［C］//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.Santiago，2020:1006-1014.
[7]Zhou K，Zhou Y，Zhao W X，et al.Towards topic-guided conversational recommender system［C］//Proceedings of the 28th International Conference on Computational Linguistics.Barcelona，2020:4128-4139.
[8]Hayati S A，Kang D，Zhu Q，et al.Inspired:toward sociable recommendation dialog systems［C］//Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing(EMNLP).Punta Cana，2020:8142-8152.
[9]Liu Z，Wang H，Niu Z Y，et al.Towards conversational recommendation over multi-type dialogs［C］//Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics(ACL).Seattle，2020:1036-1049.
[10]Moon S，Shah P，Kumar A，et al.Opendialkg:explainable conversational reasoning with attention-based walks over knowledge graphs［C］//Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics(ACL).Florence，2019:845-854.
[11]Chen H，Liu X，Yin D，et al.A survey on dialogue systems:recent advances and new frontiers［J］.ACM SIGKDD Explorations Newsletter，2017，19(2):25-35.
[12]Gao C，Lei W，He X，et al.Advances and challenges in conversational recommender systems:a survey［J］.AI Open，2021，2(1):100-126.
[13]Ravuri S，Stolcke A.Recurrent neural network and LSTM models for lexical utterance classification［C］//Proceedings of 6th Annual Conference of the International Speech Communication Association.Dresden，2015:2049-2061.
[14]Guo J，Shuang K，Li J，et al.Dual slot selector via local reliability verification for dialogue state tracking［C］//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing.Bangkok，2021:139-151.
[15]Colombo P，Chapuis E，Manica M，et al.Guiding attention in sequence-to-sequence models for dialogue act prediction［C］//Proceedings of the AAAI Conference on Artificial Intelligence.New York，2020:7594-7601.
[16]Cui Y，Che W，Liu T，et al.Pre-training with whole word masking for chinese bert［J］.IEEE/ACM Transactions on Audio，Speech，and Language Processing，2021，29(1):3504-3514.
[17]Qin L，Che W，Li Y，et al.Dcr-net:a deep co-interactive relation network for joint dialog act recognition and sentiment classification［C］//Proceedings of the AAAI Conference on Artificial Intelligence.Seattle，2020:8665-8672.
[18]Radford A，Wu J，Child R，et al.Language models are unsupervised multitask learners［J］.OpenAI Blog，2019，1(8):9-17.
[19]Vaswani A，Shazeer N，Parmar N，et al.Attention is all you need［C］//Proceedings of the 31st International Conference on Neural Information Processing Systems.Long Beach，2017:6000-6010.
[20]Zhou K，Wang X，Zhou Y，et al.CrsLab:an open-source toolkit for building conversational recommender system［C］//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing:System Demonstrations.Bangkok，2021:185-193.
[21]Gupta D，Singh K，Chakrabarti S，et al.Multi-task learning for target-dependent sentiment classification［C］//Proceedings of the Pacific-Asia Conference on Knowledge Discovery and Data Mining.Berlin:Springer，2019:185-197.
[22]Feng Y，Wang Y，Li H.A sequence-to-sequence approach to dialogue state tracking［C］//Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing.Bangkok，2021:1714-1725.

利用对话模型引导的对话生成推荐

Using Dialogue Model to Guide Recommendation of Dialogue Generation

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	耿蓉，吴亚倩，肖倩倩，徐赛. 基于改进GRU算法的天基信息网资源预测研究[J]. 东北大学学报（自然科学版）, 2023, 44(3): 305-314.
[2]	罗忠，吴东泽，李雷，葛长闯. 转子系统动力学仿真平台设计与实验验证[J]. 东北大学学报（自然科学版）, 2023, 44(3): 375-381.
[3]	侯振隆，郑玉君，巩恩普，程浩. 基于深度加权的重力梯度数据联合相关成像反演[J]. 东北大学学报:自然科学版, 2020, 41(11): 1628-1632.
[4]	吴超，李元辉，徐帅，戴星航. 地下工程岩体节理迹长统计下限值的确定方法[J]. 东北大学学报:自然科学版, 2020, 41(8): 1167-1173.
[5]	温守钦，唐铁乔，谢伟，付建飞. 氧、硫逸度对岫岩红旗铅锌矿床矿物组合共生分异的制约[J]. 东北大学学报:自然科学版, 2020, 41(7): 999-1007.
[6]	齐锡晶，唐梁，康伟鑫，秦娇娇. 高速公路桥梁桥面板养护方案多目标决策方法[J]. 东北大学学报:自然科学版, 2020, 41(7): 1033-1040.
[7]	蔡志辉，张德良，周彦君，文光奇. 应变速率对Fe-11Mn-2Al-0.2C中锰钢变形行为的影响[J]. 东北大学学报:自然科学版, 2020, 41(7): 1041-1047.
[8]	马树军，王霄霄. 吸附微粒对微悬空桥传感器的影响[J]. 东北大学学报:自然科学版, 2020, 41(1): 108-112.
[9]	马树军，玄航，孙嘉蔚，杨磊. 在非黏性流体域中微悬臂梁振动的建模与仿真[J]. 东北大学学报:自然科学版, 2018, 39(9): 1272-1276.
[10]	马树军，修强. 考虑表面应力的微悬臂梁质量传感器的建模与仿真[J]. 东北大学学报:自然科学版, 2018, 39(2): 237-241.
[11]	陈大力，沈岩涛，谢槟竹，马颖异. 基于余弦相似度模型的最佳教练遴选算法[J]. 东北大学学报:自然科学版, 2014, 35(12): 1697-1700.
[12]	潘峰，刘禄，薛定宇. 最优分数阶PIλDμ网络时延控制器[J]. 东北大学学报:自然科学版, 2014, 35(10): 1382-1385.
[13]	刘禄，潘峰，薛定宇. 一种分数阶卡尔曼滤波器[J]. 东北大学学报:自然科学版, 2014, 35(8): 1069-1072.
[14]	刘金海，冯健，马大中. 流体管道压力信号的高精度实时滤波方法[J]. 东北大学学报:自然科学版, 2013, 34(1): 9-12.
[15]	孙艳蕊. 带时间和成本约束的随机流网络可靠度的计算[J]. 东北大学学报（自然科学版）, 2013, 34(11): 1537-1541.