MIND: A Large-scale Dataset for News Recommendation

Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 3597–3606 July 5 - 10, 2020. c⃝2020 Association for Computational Linguistics 3597 MIND: A Large-scale Dataset for News Recommendation Fangzhao Wu†, Ying Qiao‡, Jiun-Hung Chen‡, Chuhan Wu§, Tao Qi§, Jianxun Lian†, Danyang Liu†, Xing Xie†, Jianfeng Gao†, Winnie Wu‡, Ming Zhou† †Microsoft Research ‡Microsoft §Tsinghua University {fangzwu, yiqia, jiuche, jialia}@microsoft.com {t-danliu, xingx, jfgao, winniew, mingzhou}@microsoft.com {wu-ch19, qit16}@mails.tsinghua.edu.cn Abstract News recommendation is an important tech- nique for personalized news service. Com- pared with product and movie recommenda- tions which have been comprehensively stud- ied, the research on news recommendation is much more limited, mainly due to the lack of a high-quality benchmark dataset. In this paper, we present a large-scale dataset named MIND for news recommendation. Constructed from the user click logs of Microsoft News, MIND contains 1 million users and more than 160k English news articles, each of which has rich textual content such as title, abstract and body. We demonstrate MIND a good testbed for news recommendation through a comparative study of several state-of-the-art news recom- mendation methods which are originally de- veloped on different proprietary datasets. Our results show the performance of news recom- mendation highly relies on the quality of news content understanding and user interest model- ing. Many natural language processing tech- niques such as effective text representation methods and pre-trained language models can effectively improve the performance of news recommendation. The MIND dataset will be available at https://msnews.github.io. 1 Introduction Online news services such as Google News and Mi- crosoft News have become important platforms for a large population of users to obtain news informa- tion (Das et al., 2007; Wu et al., 2019a). Massive news articles are generated and posted online every day, making it difficult for users to find interested news quickly (Okura et al., 2017). Personalized news recommendation can help users alleviate in- formation overload and improve news reading ex- perience (Wu et al., 2019b). Thus, it is widely used in many online news platforms (Li et al., 2011; Okura et al., 2017; An et al., 2019). In traditional recommender systems, users and items are usually represented using IDs, and their interactions such as rating scores are used to learn ID representations via methods like collaborative filtering (Koren, 2008). However, news recommen- dation has some special challenges. First, news articles on news websites update very quickly. New news articles are posted continuously, and existing news articles will expire in short time (Das et al., 2007). Thus, the cold-start problem is very severe in news recommendation. Second, news articles contain rich textual information such as title and body. It is not appropriate to simply represent- ing them using IDs, and it is important to under- stand their content from their texts (Kompan and Bielikov´a, 2010). Third, there is no explicit rating of news articles posted by users on news platforms. Thus, in news recommendation users’ interest in news is usually inferred from their click behaviors in an implicit way (Ilievski …

Content truncated to 500 words