{"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/causal-embeddings-for-recommendation","title":"Causal Embeddings for Recommendation","arxiv_id":"1706.07639","date":"2017-06-23","proceeding":null,"authors":["Stephen Bonner","Flavian vasile"],"abstract":"Many current applications use recommendations in order to modify the natural\nuser behavior, such as to increase the number of sales or the time spent on a\nwebsite. This results in a gap between the final recommendation objective and\nthe classical setup where recommendation candidates are evaluated by their\ncoherence with past user behavior, by predicting either the missing entries in\nthe user-item matrix, or the most likely next event. To bridge this gap, we\noptimize a recommendation policy for the task of increasing the desired outcome\nversus the organic user behavior. We show this is equivalent to learning to\npredict recommendation outcomes under a fully random recommendation policy. To\nthis end, we propose a new domain adaptation algorithm that learns from logged\ndata containing outcomes from a biased recommendation policy and predicts\nrecommendation outcomes according to random exposure. We compare our method\nagainst state-of-the-art factorization methods and new approaches of causal\nrecommendation and show significant improvements.","url_abs":"http://arxiv.org/abs/1706.07639v5","url_pdf":"http://arxiv.org/pdf/1706.07639v5.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":"causal-embeddings-for-recommendation","repo_url":"https://github.com/criteo-research/CausE","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1706.07639","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}