{"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/communication-algorithms-via-deep-learning","title":"Communication Algorithms via Deep Learning","arxiv_id":"1805.09317","date":"2018-05-23","proceeding":"ICLR 2018 1","authors":["Hyeji Kim","Yihan Jiang","Ranvir Rana","Sreeram Kannan","Sewoong Oh","Pramod Viswanath"],"abstract":"Coding theory is a central discipline underpinning wireline and wireless\nmodems that are the workhorses of the information age. Progress in coding\ntheory is largely driven by individual human ingenuity with sporadic\nbreakthroughs over the past century. In this paper we study whether it is\npossible to automate the discovery of decoding algorithms via deep learning. We\nstudy a family of sequential codes parameterized by recurrent neural network\n(RNN) architectures. We show that creatively designed and trained RNN\narchitectures can decode well known sequential codes such as the convolutional\nand turbo codes with close to optimal performance on the additive white\nGaussian noise (AWGN) channel, which itself is achieved by breakthrough\nalgorithms of our times (Viterbi and BCJR decoders, representing dynamic\nprograming and forward-backward algorithms). We show strong generalizations,\ni.e., we train at a specific signal to noise ratio and block length but test at\na wide range of these quantities, as well as robustness and adaptivity to\ndeviations from the AWGN setting.","url_abs":"http://arxiv.org/abs/1805.09317v1","url_pdf":"http://arxiv.org/pdf/1805.09317v1.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":"communication-algorithms-via-deep-learning","repo_url":"https://github.com/yihanjiang/Sequential-RNN-Decoder","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"communication-algorithms-via-deep-learning","repo_url":"https://github.com/TCIPrime/turbo-nn","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"communication-algorithms-via-deep-learning","repo_url":"https://github.com/datlife/deepcom","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"deep-learning","task_name":"Deep Learning"},{"task_slug":"ingenuity","task_name":"Ingenuity"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1805.09317","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}