Personalized news recommendation has been widely adopted to improve user experience. Recently, pre-trained language models (PLMs) have demonstrated the great capability of natural language understanding and the potential of improving news modeling for news recommendation. However, existing PLMs are usually pre-trained on general corpus such as BookCorpus and Wikipedia, which have some gaps with the news domain. Directly finetuning PLMs with the news recommendation task may be sub-optimal for news understanding. Besides, PLMs usually contain a large volume of parameters and have high computational overhead, which imposes a great burden on the low-latency online services. In this paper, we propose Tiny-NewsRec, which can improve both the effectiveness and the efficiency of PLM-based news recommendation. In order to reduce the domain gap between general corpora and the news data, we propose a self-supervised domain-specific post-training method to adapt the generally pre-trained language models to the news domain with the task of news title and news body matching. To improve the efficiency of PLM-based news recommendation while maintaining the performance, we propose a two-stage knowledge distillation method. In the first stage, we use the domain-specific teacher PLM to guide the student model for news semantic modeling. In the second stage, we use a multi-teacher knowledge distillation framework to transfer the comprehensive knowledge from a set of teacher models finetuned for news recommendation to the student. Experiments on two real-world datasets show that our methods can achieve better performance in news recommendation with smaller models.<\/p>\n","protected":false},"excerpt":{"rendered":"
Personalized news recommendation has been widely adopted to improve user experience. Recently, pre-trained language models (PLMs) have demonstrated the great capability of natural language understanding and the potential of improving news modeling for news recommendation. However, existing PLMs are usually pre-trained on general corpus such as BookCorpus and Wikipedia, which have some gaps with the […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","footnotes":""},"msr-content-type":[3],"msr-research-highlight":[],"research-area":[13556,13545,13555],"msr-publication-type":[193716],"msr-product-type":[],"msr-focus-area":[],"msr-platform":[],"msr-download-source":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[246691,256303,248503,248353,256189,248323,257923,256555,255229,251989],"msr-conference":[260143],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-900936","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-research-area-human-language-technologies","msr-research-area-search-information-retrieval","msr-locale-en_us","msr-field-of-study-computer-science","msr-field-of-study-domain-software-engineering","msr-field-of-study-information-retrieval","msr-field-of-study-language-model","msr-field-of-study-matching-statistics","msr-field-of-study-natural-language-understanding","msr-field-of-study-order-exchange","msr-field-of-study-set-psychology","msr-field-of-study-task-project-management","msr-field-of-study-user-experience-design"],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2021-12-1","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/arxiv.org\/abs\/2112.00944","label_id":"243109","label":0}],"msr_related_uploader":"","msr_attachments":[],"msr-author-ordering":[{"type":"guest","value":"yang-yu-2","user_id":900942,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=yang-yu-2"},{"type":"user_nicename","value":"Fangzhao Wu","user_id":36473,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Fangzhao Wu"},{"type":"guest","value":"chuhan-wu","user_id":782656,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=chuhan-wu"},{"type":"guest","value":"jingwei-yi","user_id":900948,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=jingwei-yi"},{"type":"guest","value":"tao-qi","user_id":782662,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=tao-qi"},{"type":"guest","value":"qi-liu-2","user_id":900951,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=qi-liu-2"}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[886398],"msr_group":[],"msr_project":[],"publication":[],"video":[],"download":[],"msr_publication_type":"inproceedings","related_content":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/900936"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":1,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/900936\/revisions"}],"predecessor-version":[{"id":900963,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/900936\/revisions\/900963"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=900936"}],"wp:term":[{"taxonomy":"msr-content-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-content-type?post=900936"},{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=900936"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=900936"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=900936"},{"taxonomy":"msr-product-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-product-type?post=900936"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=900936"},{"taxonomy":"msr-platform","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-platform?post=900936"},{"taxonomy":"msr-download-source","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-download-source?post=900936"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=900936"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=900936"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=900936"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=900936"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=900936"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=900936"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=900936"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}