{"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/tensor-train-recurrent-neural-networks-for","title":"Tensor-Train Recurrent Neural Networks for Video Classification","arxiv_id":"1707.01786","date":"2017-07-06","proceeding":"ICML 2017 8","authors":["Yinchong Yang","Denis Krompass","Volker Tresp"],"abstract":"The Recurrent Neural Networks and their variants have shown promising\nperformances in sequence modeling tasks such as Natural Language Processing.\nThese models, however, turn out to be impractical and difficult to train when\nexposed to very high-dimensional inputs due to the large input-to-hidden weight\nmatrix. This may have prevented RNNs' large-scale application in tasks that\ninvolve very high input dimensions such as video modeling; current approaches\nreduce the input dimensions using various feature extractors. To address this\nchallenge, we propose a new, more general and efficient approach by factorizing\nthe input-to-hidden weight matrix using Tensor-Train decomposition which is\ntrained simultaneously with the weights themselves. We test our model on\nclassification tasks using multiple real-world video datasets and achieve\ncompetitive performances with state-of-the-art models, even though our model\narchitecture is orders of magnitude less complex. We believe that the proposed\napproach provides a novel and fundamental building block for modeling\nhigh-dimensional sequential data with RNN architectures and opens up many\npossibilities to transfer the expressive and advanced architectures from other\ndomains such as NLP to modeling high-dimensional sequential data.","url_abs":"http://arxiv.org/abs/1707.01786v1","url_pdf":"http://arxiv.org/pdf/1707.01786v1.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":"tensor-train-recurrent-neural-networks-for","repo_url":"https://github.com/Tuyki/TT_RNN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"video-classification","task_name":"Video Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1707.01786","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}