{"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/structured-semantic-model-supported-deep","title":"Structured Semantic Model supported Deep Neural Network for Click-Through Rate Prediction","arxiv_id":"1812.01353","date":"2018-12-04","proceeding":null,"authors":["Chenglei Niu","Guojing Zhong","Ying Liu","Yandong Zhang","Yongsheng Sun","Ailong He","Zhaoji Chen"],"abstract":"With the rapid development of online advertising and recommendation systems,\nclick-through rate prediction is expected to play an increasingly important\nrole.Recently many DNN-based models which follow a similar Embedding&MLP\nparadigm have been proposed, and have achieved good result in image/voice and\nnlp fields. In these methods the Wide&Deep model announced by Google plays a\nkey role.Most models first map large scale sparse input features into\nlow-dimensional vectors which are transformed to fixed-length vectors, then\nconcatenated together before being fed into a multilayer perceptron (MLP) to\nlearn non-linear relations among input features. The number of trainable\nvariables normally grow dramatically the number of feature fields and the\nembedding dimension grow. It is a big challenge to get state-of-the-art result\nthrough training deep neural network and embedding together, which falls into\nlocal optimal or overfitting easily. In this paper, we propose an Structured\nSemantic Model (SSM) to tackles this challenge by designing a orthogonal base\nconvolution and pooling model which adaptively learn the multi-scale base\nsemantic representation between features supervised by the click label.The\noutput of SSM are then used in the Wide&Deep for CTR prediction.Experiments on\ntwo public datasets as well as real Weibo production dataset with over 1\nbillion samples have demonstrated the effectiveness of our proposed approach\nwith superior performance comparing to state-of-the-art methods.","url_abs":"http://arxiv.org/abs/1812.01353v5","url_pdf":"http://arxiv.org/pdf/1812.01353v5.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":"structured-semantic-model-supported-deep","repo_url":"https://github.com/niuchenglei/ssm-dnn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"structured-semantic-model-supported-deep","repo_url":"https://github.com/niuchenglei/usm-dnn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"click-through-rate-prediction","task_name":"Click-Through Rate Prediction"},{"task_slug":"recommendation-systems","task_name":"Recommendation Systems"}],"methods":[{"method_slug":"wide-deep","method_name":"Wide&Deep"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}