{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/interaction-aware-factorization-machines-for","title":"Interaction-aware Factorization Machines for Recommender Systems","arxiv_id":"1902.09757","date":"2019-02-26","proceeding":null,"authors":["Fuxing Hong","Dongbo Huang","Ge Chen"],"abstract":"Factorization Machine (FM) is a widely used supervised learning approach by\neffectively modeling of feature interactions. Despite the successful\napplication of FM and its many deep learning variants, treating every feature\ninteraction fairly may degrade the performance. For example, the interactions\nof a useless feature may introduce noises; the importance of a feature may also\ndiffer when interacting with different features. In this work, we propose a\nnovel model named \\emph{Interaction-aware Factorization Machine} (IFM) by\nintroducing Interaction-Aware Mechanism (IAM), which comprises the\n\\emph{feature aspect} and the \\emph{field aspect}, to learn flexible\ninteractions on two levels. The feature aspect learns feature interaction\nimportance via an attention network while the field aspect learns the feature\ninteraction effect as a parametric similarity of the feature interaction vector\nand the corresponding field interaction prototype. IFM introduces more\nstructured control and learns feature interaction importance in a stratified\nmanner, which allows for more leverage in tweaking the interactions on both\nfeature-wise and field-wise levels. Besides, we give a more generalized\narchitecture and propose Interaction-aware Neural Network (INN) and DeepIFM to\ncapture higher-order interactions. To further improve both the performance and\nefficiency of IFM, a sampling scheme is developed to select interactions based\non the field aspect importance. The experimental results from two well-known\ndatasets show the superiority of the proposed models over the state-of-the-art\nmethods.","url_abs":"http://arxiv.org/abs/1902.09757v1","url_pdf":"http://arxiv.org/pdf/1902.09757v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"interaction-aware-factorization-machines-for","repo_url":"https://github.com/cstur4/interaction-aware-factorization-machines","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"recommendation-systems","task_name":"Recommendation Systems"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/recommendation-systems-on-frappe","task":"Recommendation Systems","dataset":"Frappe","model":"INN","rank_in_archive_order":1,"of":2,"metrics":{"RMSE":"0.3071"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}