Kate Knill

Kate Knill is a speech researcher whose work spans multilingual speech recognition, speech synthesis and assessment for English learners. She earned a first-class engineering and mathematics degree at Nottingham and a PhD in digital signal processing at Imperial, researched audio document retrieval at Cambridge, and developed speech systems at Nuance and Toshiba. At Cambridge, she contributed to Speak & Improve and served ISCA. She coauthored the 2025 Speak & Improve Challenge and Corpus, a 2026 study of bias in automated speech assessment, and gave a Division F keynote on English learner speech. As of October 2026, she is a Principal Research Associate in Cambridge's Machine Intelligence Laboratory and a principal investigator for ALTA's Spoken Language Technology Project.

Now at: University of Cambridge
Badges: Bridge

Timeline

1990–1996

Degrees and speech retrieval

She earns a first-class BEng in electrical engineering and mathematics at Nottingham in 1990 and a PhD in adaptive digital signal processing at Imperial College in 1994; Marconi Underwater Systems sponsors both degrees and her pre-university year. From 1993 to 1996, she works as a Research Associate in Cambridge's Speech Vision and Robotics Group on audio document retrieval, supervised by Steve Young and funded by HP Labs in Bristol.
1997–2002

Multilingual systems at Nuance

She joins Nuance Communications' speech R&D team in 1997 and leads a cross-site team as Languages Manager from 2000 to 2002. The team develops speech recognition and speaker verification in more than 20 languages.
2002–2012

Speech Technology Group at Toshiba

In 2002, she establishes Toshiba Research Europe's Speech Technology Group at its Cambridge Research Laboratory and leads work on speech-recognition and speech-synthesis engines.
2012–2021

Cambridge research and ISCA service

In July 2012, she leaves Toshiba's Cambridge Research Lab to become a Senior Research Associate at the University of Cambridge. By 2019, she works in Cambridge's Machine Intelligence Laboratory and contributes to Speak & Improve, a system that helps English learners understand their speaking ability for work, study and exams. She serves on ISCA's Board from 2013 to 2021 and as Secretary in its 2017–2019 and 2019–2021 terms.
2025

The Speak & Improve Challenge

At the 2025 SLaTE workshop in Nijmegen, she serves as a general chair and coauthors the Speak & Improve Challenge 2025. Its four shared tasks cover automatic speech recognition, spoken-language assessment, spoken grammatical-error correction and feedback.
2025

The Speak & Improve Corpus

She coauthors the Speak & Improve Corpus 2025, a dataset of L2 English speech with holistic scores and language-error annotations. It supports research on low-resource L2 speech recognition and spoken grammatical-error correction.
August 2026

Bias in speaking assessment

In August 2026, she coauthors a study of how speaking proficiency, first language and age relate to automated L2 speaking-assessment scores. It analyzes a BERT text grader and a Whisper-based speech-and-text grader.
As of October 2026

Principal Research Associate at Cambridge

As of October 2026, she is a Principal Research Associate in Cambridge's Machine Intelligence Laboratory and a Principal Investigator for the ALTA Institute Spoken Language Technology Project. In 2026, she gives a Division F keynote on automated assessment of second-language English learners' speech.