{"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/first-pass-large-vocabulary-continuous-speech","title":"First-Pass Large Vocabulary Continuous Speech Recognition using Bi-Directional Recurrent DNNs","arxiv_id":"1408.2873","date":"2014-08-12","proceeding":null,"authors":["Awni Y. Hannun","Andrew L. Maas","Daniel Jurafsky","Andrew Y. Ng"],"abstract":"We present a method to perform first-pass large vocabulary continuous speech\nrecognition using only a neural network and language model. Deep neural network\nacoustic models are now commonplace in HMM-based speech recognition systems,\nbut building such systems is a complex, domain-specific task. Recent work\ndemonstrated the feasibility of discarding the HMM sequence modeling framework\nby directly predicting transcript text from audio. This paper extends this\napproach in two ways. First, we demonstrate that a straightforward recurrent\nneural network architecture can achieve a high level of accuracy. Second, we\npropose and evaluate a modified prefix-search decoding algorithm. This approach\nto decoding enables first-pass speech recognition with a language model,\ncompletely unaided by the cumbersome infrastructure of HMM-based systems.\nExperiments on the Wall Street Journal corpus demonstrate fairly competitive\nword error rates, and the importance of bi-directional network recurrence.","url_abs":"http://arxiv.org/abs/1408.2873v2","url_pdf":"http://arxiv.org/pdf/1408.2873v2.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":"first-pass-large-vocabulary-continuous-speech","repo_url":"https://github.com/baidu-research/warp-ctc","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"torch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"first-pass-large-vocabulary-continuous-speech","repo_url":"https://github.com/jb1999/eesen","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"first-pass-large-vocabulary-continuous-speech","repo_url":"https://github.com/srvk/eesen","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"first-pass-large-vocabulary-continuous-speech","repo_url":"https://github.com/taozitongxue1/11bee-DeepSpeech","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"first-pass-large-vocabulary-continuous-speech","repo_url":"https://github.com/PaddlePaddle/PaddleSpeech","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"paddle","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"speech-recognition","task_name":"Speech Recognition"},{"task_slug":"speech-recognition-1","task_name":"speech-recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1408.2873","atlas_url":"https://app.syntology.ai/?focus=1408.2873","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}