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
  • [1] An MX, 2019, 57TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2019), P336
  • [2] [Anonymous], 2021, P 2021 ACM SIGIR INT, DOI DOI 10.1109/IROS51168.2021.9636361
  • [3] Bahdanau D, 2016, Arxiv, DOI [arXiv:1409.0473, 10.48550/arXiv.1409.0473]
  • [4] Bordes Antoine, 2013, P 26 INT C NEUR INF, V26
  • [5] Chih-Ming Wu, 2020, 2020 IEEE Eurasia Conference on IOT, Communication and Engineering (ECICE), P44, DOI 10.1109/ECICE50847.2020.9301944
  • [6] Cho K., 2014, P 2014 C EMP METH NA, P1724, DOI 10.3115/v1/d14-1179
  • [7] Positive, Negative and Neutral: Modeling Implicit Feedback in Session-based News Recommendation
    Gong, Shansan
    Zhu, Kenny Q.
    [J]. PROCEEDINGS OF THE 45TH INTERNATIONAL ACM SIGIR CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION RETRIEVAL (SIGIR '22), 2022, : 1185 - 1195
  • [8] GULLA JA, 2017, P INT C WEB INT, P1042, DOI DOI 10.1145/3106426.3109436
  • [9] Huang PS, 2013, PROCEEDINGS OF THE 22ND ACM INTERNATIONAL CONFERENCE ON INFORMATION & KNOWLEDGE MANAGEMENT (CIKM'13), P2333
  • [10] Iana Andreea, 2022, SEMANTIC WEB PREPRIN, P1