{"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/the-ibm-2016-english-conversational-telephone","title":"The IBM 2016 English Conversational Telephone Speech Recognition System","arxiv_id":"1604.08242","date":"2016-04-27","proceeding":null,"authors":["George Saon","Tom Sercu","Steven Rennie","Hong-Kwang J. Kuo"],"abstract":"We describe a collection of acoustic and language modeling techniques that\nlowered the word error rate of our English conversational telephone LVCSR\nsystem to a record 6.6% on the Switchboard subset of the Hub5 2000 evaluation\ntestset. On the acoustic side, we use a score fusion of three strong models:\nrecurrent nets with maxout activations, very deep convolutional nets with 3x3\nkernels, and bidirectional long short-term memory nets which operate on FMLLR\nand i-vector features. On the language modeling side, we use an updated model\n\"M\" and hierarchical neural network LMs.","url_abs":"http://arxiv.org/abs/1604.08242v2","url_pdf":"http://arxiv.org/pdf/1604.08242v2.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":[],"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":[{"method_slug":"maxout","method_name":"Maxout"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/speech-recognition-on-switchboard-hub500","task":"Speech Recognition","dataset":"Switchboard + Hub500","model":"RNN + VGG + LSTM acoustic model trained on SWB+Fisher+CH, N-gram + \"model M\" + NNLM language model","rank_in_archive_order":7,"of":30,"metrics":{"Percentage error":"6.6"},"uses_additional_data":false},{"leaderboard":"/sota/speech-recognition-on-switchboard-hub500","task":"Speech Recognition","dataset":"Switchboard + Hub500","model":"IBM 2016","rank_in_archive_order":9,"of":30,"metrics":{"Percentage error":"6.9"},"uses_additional_data":false},{"leaderboard":"/sota/speech-recognition-on-swb_hub_500-wer","task":"Speech Recognition","dataset":"swb_hub_500 WER fullSWBCH","model":"RNN + VGG + LSTM acoustic model trained on SWB+Fisher+CH, N-gram + \"model M\" + NNLM language model","rank_in_archive_order":5,"of":12,"metrics":{"Percentage error":"12.2"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}