--- title: "Adam Coates — author profile and papers" description: "Adam Coates, University of Graz: 2 papers on Elamur Research (2014–2015), h-index 29 — affiliations, topics, co-authors and who cites them in our library." canonical: "https://research.elamur.ai/free/en/authors/adam-coates/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/adam-coates/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Adam Coates # Adam Coates Adam Coates works on automatic [speech recognition](https://md.research.elamur.ai/free/en/topics/speech/). The papers here, from 2014 and 2015, ask whether a single neural network can do the job of a traditional recognizer, which is normally assembled from separately tuned feature extractors, acoustic and pronunciation models, and language models. The approach is end-to-end training with the CTC loss, so audio needs no prior alignment to transcripts. The work relies on large amounts of transcribed speech and on engineering that spreads training across many GPUs. The earlier of the two papers, [Deep Speech: Scaling up end-to-end speech recognition](https://md.research.elamur.ai/free/en/papers/hannun-2014-deep-speech-scaling-up-end/) (2014), maps spectrograms directly to English characters. Its recurrent model is kept deliberately simple, with one bidirectional recurrent layer and no LSTM gates, so that it runs efficiently in parallel. It was trained on thousands of hours of speech and is designed to cope better with noisy input. The follow-up, [Deep Speech 2](https://md.research.elamur.ai/free/en/papers/amodei-2015-deep-speech-2-end-end/) (2015), carries the same idea to both English and Mandarin. It combines convolutional and recurrent layers, applies batch normalization to the recurrent part, and adds synthetic augmentation to roughly twelve thousand hours of English and over nine thousand hours of Mandarin audio. The paper also uses synchronous SGD on a high-performance computing stack, which makes training about seven times faster than the first system. Other contributions are human transcription of the test sets, so that recognition accuracy can be compared with measured human performance, and a pruned beam search that speeds up decoding. The bylines of these papers place Coates at Stanford University and later at Baidu, where both Deep Speech systems were built. Frequent co-authors include [Bryan Catanzaro](https://md.research.elamur.ai/free/en/authors/bryan-catanzaro/), [Greg Diamos](https://md.research.elamur.ai/free/en/authors/greg-diamos/) and [Erich Elsen](https://md.research.elamur.ai/free/en/authors/erich-elsen/). The Deep Speech 2 team also included [Dario Amodei](https://md.research.elamur.ai/free/en/authors/dario-amodei/) and [Andrew Y. Ng](https://md.research.elamur.ai/free/en/authors/andrew-y-ng/). 2 papers on this site · 2014–2015 current affiliation [University of Graz](https://ror.org/01faaaf77), Graz, Austria; [BioTechMed-Graz](https://ror.org/02jfbm483), Graz, Austria Adam Coates has 2 papers on this site, in 2014–2015, most of them on [Speech & spoken language](https://md.research.elamur.ai/free/en/topics/speech/); often with [Bryan Catanzaro](https://md.research.elamur.ai/free/en/authors/bryan-catanzaro/), [Carl Case](https://md.research.elamur.ai/free/en/authors/carl-case/), [Erich Elsen](https://md.research.elamur.ai/free/en/authors/erich-elsen/). [OpenAlex](https://openalex.org/A5103153292) · [ORCID](https://orcid.org/0000-0002-4004-1603) · [**h-index 29**](https://openalex.org/authors/A5103153292) · [49 works](https://openalex.org/works?filter=author.id:A5103153292) · [13 801 citations](https://openalex.org/works?filter=author.id:A5103153292&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [2 works on this site](https://md.research.elamur.ai/free/en/authors/adam-coates/#papers) · [331 citations in our library](https://md.research.elamur.ai/free/en/authors/adam-coates/cited-by/) The profile is the library's match of the name; two people with one name are not always told apart. Elsewhere: [Google Scholar](https://scholar.google.com/citations?view_op=search_authors&mauthors=Adam+Coates) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Adam+Coates) · [dblp](https://dblp.org/search?q=Adam+Coates) · [ACL Anthology](https://aclanthology.org/search/?q=Adam+Coates) ## Affiliations - [BioTechMed-Graz](https://ror.org/02jfbm483) · Graz, Austria —current · 2024–2026 on 2 papers - [University of Graz](https://ror.org/01faaaf77) · Graz, Austria —current · 2024–2026 on 2 papers - [Manchester Metropolitan University](https://ror.org/02hstj355) · Manchester, United Kingdom —2017–2018 on 2 papers - [Stanford University](https://md.research.elamur.ai/free/en/universities/stanford-university/) · Stanford, United States —2005–2017 on 26 papers - [Baidu (China)](https://md.research.elamur.ai/free/en/companies/baidu/) · Beijing, China —2016–2017 on 4 papers From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 2 more institutions on single bylines are not listed. ## Topics [Speech & spoken language 2](https://md.research.elamur.ai/free/en/topics/speech/) ## Works by year ### 2010s [**2015** 1 paper](https://md.research.elamur.ai/free/en/papers/amodei-2015-deep-speech-2-end-end/) · [**2014** 1 paper](https://md.research.elamur.ai/free/en/papers/hannun-2014-deep-speech-scaling-up-end/) ## Papers The most cited first (OpenAlex and our library). 1. **[Deep Speech 2: End-to-End Speech Recognition in English and Mandarin](https://md.research.elamur.ai/free/en/papers/amodei-2015-deep-speech-2-end-end/)** *(2015 · 2 302 citations)* Conventional speech recognizers are assembled from many hand-tuned parts, including feature extractors, acoustic, pronunciation and language models, and speaker adaptation. 2. **[Deep Speech: Scaling up end-to-end speech recognition](https://md.research.elamur.ai/free/en/papers/hannun-2014-deep-speech-scaling-up-end/)** *(2014 · 1 575 citations)* Conventional speech recognizers chain together many hand-tuned components, including engineered features, acoustic models, HMMs and phoneme dictionaries. ## Co-authors here 10 co-authors with a page, the most papers together first. - [Bryan Catanzaro](https://md.research.elamur.ai/free/en/authors/bryan-catanzaro/) ×2 - [Carl Case](https://md.research.elamur.ai/free/en/authors/carl-case/) ×2 - [Erich Elsen](https://md.research.elamur.ai/free/en/authors/erich-elsen/) ×2 - [Greg Diamos](https://md.research.elamur.ai/free/en/authors/greg-diamos/) ×2 - [Jared Casper](https://md.research.elamur.ai/free/en/authors/jared-casper/) ×2 - [Andrew Y. Ng](https://md.research.elamur.ai/free/en/authors/andrew-y-ng/) - [Dario Amodei](https://md.research.elamur.ai/free/en/authors/dario-amodei/) - [Jingdong Chen](https://md.research.elamur.ai/free/en/authors/jingdong-chen/) - [Ryan Prenger](https://md.research.elamur.ai/free/en/authors/ryan-prenger/) - [Sanjeev Satheesh](https://md.research.elamur.ai/free/en/authors/sanjeev-satheesh/) ## Cited by Papers of our library cite their works 331 times (of the 140 884 papers whose bibliographies we read); [23 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/adam-coates/cited-by/).