{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T19:08:52Z","timestamp":1760123332223,"version":"3.33.0"},"reference-count":41,"publisher":"IEEE","license":[{"start":{"date-parts":[[2024,12,2]],"date-time":"2024-12-02T00:00:00Z","timestamp":1733097600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2024,12,2]],"date-time":"2024-12-02T00:00:00Z","timestamp":1733097600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,12,2]]},"DOI":"10.1109\/slt61566.2024.10832249","type":"proceedings-article","created":{"date-parts":[[2025,1,16]],"date-time":"2025-01-16T18:31:27Z","timestamp":1737052287000},"page":"102-109","source":"Crossref","is-referenced-by-count":1,"title":["Training Large ASR Encoders With Differential Privacy"],"prefix":"10.1109","author":[{"given":"Geeticka","family":"Chauhan","sequence":"first","affiliation":[{"name":"Google"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Steve","family":"Chien","sequence":"additional","affiliation":[{"name":"Google"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Om","family":"Thakkar","sequence":"additional","affiliation":[{"name":"Google"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Abhradeep","family":"Thakurta","sequence":"additional","affiliation":[{"name":"Google"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Arun","family":"Narayanan","sequence":"additional","affiliation":[{"name":"Google"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","article-title":"Attention is all you need","author":"Vaswani","year":"2017","journal-title":"Neural Information Processing Systems (NeurIPS)"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP40776.2020.9053896"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.21437\/interspeech.2019-1819"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2020-3015"},{"article-title":"Self-supervised learning with random-projection quantizer for speech recognition","volume-title":"International Conference on Machine Learning","author":"Chiu","key":"ref5"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/SP.2017.41"},{"article-title":"The secret sharer: Evaluating and testing unintended memorization in neural networks","volume-title":"28th USENIX Security Symposium","author":"Carlini","key":"ref7"},{"article-title":"Extracting training data from large language models","volume-title":"30th USENIX Security Symposium","author":"Carlini","key":"ref8"},{"article-title":"Extracting training data from diffusion models","volume-title":"USENIX Security Symposium","author":"Carlini","key":"ref9"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2022-10895"},{"article-title":"Measuring forgetting of memorized training examples","volume-title":"The International Conference on Learning Representations (ICLR)","author":"Jagielski","key":"ref11"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP48485.2024.10446083"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP48485.2024.10447773"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2024-574"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/1143844.1143891"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1007\/11681878_14"},{"key":"ref17","article-title":"Practical and private (deep) learning without sampling or shuffling","author":"Kairouz","year":"2021","journal-title":"ICML"},{"key":"ref18","article-title":"Federated learning with differential privacy for end-to-end speech recognition","author":"Pelikan","year":"2023","journal-title":"arXiv preprint arXiv:2310.00098"},{"key":"ref19","article-title":"Differentially private fine-tuning of language models","author":"Yu","year":"2022","journal-title":"ICLR"},{"key":"ref20","article-title":"Differentially private bias-term only fine-tuning of foundation models","author":"Bu","year":"2022","journal-title":"arXiv preprint arXiv:2210.00036"},{"key":"ref21","article-title":"LoRA: Low-rank adaptation of large language models","author":"Hu","year":"2022","journal-title":"ICLR"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/FOCS.2014.56"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.findings-acl.171"},{"key":"ref24","article-title":"Learning differentially private recurrent language models","author":"Brendan McMahan","year":"2018","journal-title":"ICLR"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1561\/9781601988195"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1613\/jair.1.14649"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1145\/2976749.2978318"},{"key":"ref28","article-title":"Enabling fast differentially private sgd via just-in-time compilation and vectorization","author":"Subramani","year":"2021","journal-title":"NeurIPS"},{"key":"ref29","article-title":"When does differentially private learning not suffer in high dimensions?","author":"Li","year":"2022","journal-title":"NeurIPS"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.findings-emnlp.484"},{"key":"ref31","article-title":"Toward training at imagenet scale with differential privacy","author":"Kurakin","year":"2022","journal-title":"arXiv preprint arXiv:2201.12328"},{"key":"ref32","article-title":"Unlocking high-accuracy differentially private image classification through scale","author":"De","year":"2022","journal-title":"arXiv preprint arXiv:2204.13650"},{"key":"ref33","article-title":"Large language models can be strong differentially private learners","author":"Li","year":"2021","journal-title":"arXiv preprint arXiv:2110.05679"},{"key":"ref34","article-title":"Inference and interference: The role of clipping, pruning and loss landscapes in differentially private stochastic gradient descent","author":"Watson","year":"2023","journal-title":"arXiv preprint arXiv:2311.06839"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/JSTSP.2022.3182537"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP40776.2020.9052942"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2015.7178964"},{"key":"ref38","article-title":"R\u00e9nyi differential privacy of the sampled gaussian mechanism","author":"Mironov","year":"2019","journal-title":"arXiv preprint arXiv:1908.10530"},{"key":"ref39","article-title":"Why is public pretraining necessary for private model training?","author":"Ganesh","year":"2023","journal-title":"ICML"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR46437.2021.00502"},{"key":"ref41","article-title":"Public data-assisted mirror descent for private model training","author":"Amid","year":"2022","journal-title":"ICML"}],"event":{"name":"2024 IEEE Spoken Language Technology Workshop (SLT)","start":{"date-parts":[[2024,12,2]]},"location":"Macao","end":{"date-parts":[[2024,12,5]]}},"container-title":["2024 IEEE Spoken Language Technology Workshop (SLT)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/10830790\/10830793\/10832249.pdf?arnumber=10832249","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,1,17]],"date-time":"2025-01-17T08:17:54Z","timestamp":1737101874000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10832249\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,12,2]]},"references-count":41,"URL":"https:\/\/doi.org\/10.1109\/slt61566.2024.10832249","relation":{},"subject":[],"published":{"date-parts":[[2024,12,2]]}}}