{"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/end-to-end-learning-of-lda-by-mirror-descent","title":"End-to-end Learning of LDA by Mirror-Descent Back Propagation over a Deep Architecture","arxiv_id":"1508.03398","date":"2015-08-14","proceeding":"NeurIPS 2015 12","authors":["Jianshu Chen","Ji He","Yelong Shen","Lin Xiao","Xiaodong He","Jianfeng Gao","Xinying Song","Li Deng"],"abstract":"We develop a fully discriminative learning approach for supervised Latent\nDirichlet Allocation (LDA) model using Back Propagation (i.e., BP-sLDA), which\nmaximizes the posterior probability of the prediction variable given the input\ndocument. Different from traditional variational learning or Gibbs sampling\napproaches, the proposed learning method applies (i) the mirror descent\nalgorithm for maximum a posterior inference and (ii) back propagation over a\ndeep architecture together with stochastic gradient/mirror descent for model\nparameter estimation, leading to scalable and end-to-end discriminative\nlearning of the model. As a byproduct, we also apply this technique to develop\na new learning method for the traditional unsupervised LDA model (i.e.,\nBP-LDA). Experimental results on three real-world regression and classification\ntasks show that the proposed methods significantly outperform the previous\nsupervised topic models, neural networks, and is on par with deep neural\nnetworks.","url_abs":"http://arxiv.org/abs/1508.03398v2","url_pdf":"http://arxiv.org/pdf/1508.03398v2.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":"end-to-end-learning-of-lda-by-mirror-descent","repo_url":"https://github.com/jvking/bp-lda","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"topic-models","task_name":"Topic Models"},{"task_slug":"parameter-estimation","task_name":"parameter estimation"}],"methods":[{"method_slug":"lda","method_name":"LDA"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1508.03398","atlas_url":"https://app.syntology.ai/?focus=1508.03398","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}