Simplifying Content-Based Neural News Recommendation: On User Modeling and Training Objectives

被引:3
作者
Iana, Andreea [1 ]
Glavas, Goran [2 ]
Paulheim, Heiko [1 ]
机构
[1] Univ Mannheim, Mannheim, Germany
[2] Univ Wurzburg, CAIDAS, Wurzburg, Germany
来源
PROCEEDINGS OF THE 46TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL, SIGIR 2023 | 2023年
关键词
neural news recommendation; user modeling; late fusion; training objectives; contrastive learning; evaluation;
D O I
10.1145/3539618.3592062
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The advent of personalized news recommendation has given rise to increasingly complex recommender architectures. Most neural news recommenders rely on user click behavior and typically introduce dedicated user encoders that aggregate the content of clicked news into user embeddings (early fusion). These models are predominantly trained with standard point-wise classification objectives. The existing body of work exhibits two main shortcomings: (1) despite general design homogeneity, direct comparisons between models are hindered by varying evaluation datasets and protocols; (2) it leaves alternative model designs and training objectives vastly unexplored. In this work, we present a unified framework for news recommendation, allowing for a systematic and fair comparison of news recommenders across several crucial design dimensions: (i) candidate-awareness in user modeling, (ii) click behavior fusion, and (iii) training objectives. Our findings challenge the status quo in neural news recommendation. We show that replacing sizable user encoders with parameter-efficient dot products between candidate and clicked news embeddings (late fusion) often yields substantial performance gains. Moreover, our results render contrastive training a viable alternative to point-wise classification objectives.
引用
收藏
页码:2384 / 2388
页数:5
相关论文
共 42 条
  • [31] Is News Recommendation a Sequential Recommendation Task?
    Wu, Chuhan
    Wu, Fangzhao
    Qi, Tao
    Li, Chenliang
    Huang, Yongfeng
    [J]. PROCEEDINGS OF THE 45TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL (SIGIR '22), 2022, : 2382 - 2386
  • [32] Wu CH, 2019, 2019 CONFERENCE ON EMPIRICAL METHODS IN NATURAL LANGUAGE PROCESSING AND THE 9TH INTERNATIONAL JOINT CONFERENCE ON NATURAL LANGUAGE PROCESSING (EMNLP-IJCNLP 2019), P6389
  • [33] Empowering News Recommendation with Pre-trained Language Models
    Wu, Chuhan
    Wu, Fangzhao
    Qi, Tao
    Huang, Yongfeng
    [J]. SIGIR '21 - PROCEEDINGS OF THE 44TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL, 2021, : 1652 - 1656
  • [34] Wu CH, 2019, PROCEEDINGS OF THE TWENTY-EIGHTH INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, P3863
  • [35] Wu CH, 2019, 57TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2019), P1154
  • [36] NPA: Neural News Recommendation with Personalized Attention
    Wu, Chuhan
    Wu, Fangzhao
    An, Mingxiao
    Huang, Jianqiang
    Huang, Yongfeng
    Xie, Xing
    [J]. KDD'19: PROCEEDINGS OF THE 25TH ACM SIGKDD INTERNATIONAL CONFERENCCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2019, : 2576 - 2584
  • [37] Wu Chuhan, 2022, P 31 INT JOINT C ART, P2509, DOI DOI 10.24963/IJCAI.2022/348
  • [38] Wu FZ, 2020, 58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020), P3597
  • [39] Contrastive Learning for Sequential Recommendation
    Xie, Xu
    Sun, Fei
    Liu, Zhaoyang
    Wu, Shiwen
    Gao, Jinyang
    Zhang, Jiandong
    Ding, Bolin
    Cui, Bin
    [J]. 2022 IEEE 38TH INTERNATIONAL CONFERENCE ON DATA ENGINEERING (ICDE 2022), 2022, : 1259 - 1273
  • [40] Supervised contrastive learning for recommendation
    Yang, Chun
    Zou, Jianxiao
    Wu, JianHua
    Xu, Hongbing
    Fan, Shicai
    [J]. KNOWLEDGE-BASED SYSTEMS, 2022, 258