{"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/learning-to-adapt-a-meta-learning-approach","title":"Learning to adapt: a meta-learning approach for speaker adaptation","arxiv_id":"1808.10239","date":"2018-08-30","proceeding":null,"authors":["Ondřej Klejch","Joachim Fainberg","Peter Bell"],"abstract":"The performance of automatic speech recognition systems can be improved by\nadapting an acoustic model to compensate for the mismatch between training and\ntesting conditions, for example by adapting to unseen speakers. The success of\nspeaker adaptation methods relies on selecting weights that are suitable for\nadaptation and using good adaptation schedules to update these weights in order\nnot to overfit to the adaptation data. In this paper we investigate a\nprincipled way of adapting all the weights of the acoustic model using a\nmeta-learning. We show that the meta-learner can learn to perform supervised\nand unsupervised speaker adaptation and that it outperforms a strong baseline\nadapting LHUC parameters when adapting a DNN AM with 1.5M parameters. We also\nreport initial experiments on adapting TDNN AMs, where the meta-learner\nachieves comparable performance with LHUC.","url_abs":"http://arxiv.org/abs/1808.10239v1","url_pdf":"http://arxiv.org/pdf/1808.10239v1.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":"learning-to-adapt-a-meta-learning-approach","repo_url":"https://github.com/choko/learning_to_adapt","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"automatic-speech-recognition-2","task_name":"Automatic Speech Recognition"},{"task_slug":"automatic-speech-recognition","task_name":"Automatic Speech Recognition (ASR)"},{"task_slug":"meta-learning","task_name":"Meta-Learning"},{"task_slug":"speech-recognition","task_name":"Speech Recognition"},{"task_slug":"speech-recognition-1","task_name":"speech-recognition"}],"methods":[{"method_slug":"am","method_name":"AM"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.10239","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}