Machine Intelligence Laboratory

Cambridge University Department of Engineering

Anton Ragni

Background - Projects - Research - Publications

Position: Research Associate

E-mail: ar527 [at]

Office Location: BE4-58


I received my B.Eng. and M.Eng. degrees in information technology from University of Tartu in 2005 and 2007 respectively. From 2005 to 2008 I underwent graduate training at Nordic Graduate School of Language Technology. From 2007 to 2008 I was an intern in Speech Technology Group, Toshiba Research Europe Limited. From 2008 to 2014 I was studying for my PhD degree in automatic speech recognition at University of Cambridge, where my supervisor was Professor Mark Gales. I am now a Research Associate workin gfor Prof Gales on the BABEL project.

I received the "Best Student Paper Award" at IEEE Workshop on Automatic Speech Recognition and Understanding for my paper "Generative kernels for noise robust ASR" co-authored with M.J.F. Gales in 2011.


Currently associated with projects:

BABEL (IARPA funded)

Generative kernels and score-spaces for classification of speech (non-funded)

HTK Version 3 (non-funded)

In past associated with projects:

Discriminative models for speech (Toshiba, EPSRC and HTK funded PhD)

HTK Version 3

In past collaborated on projects:

Kernel Methods for Speech Processing (non-funded) with Z. Roupakia and M.J.F. Gales

Structured Discriminative Models (non-funded) with S.-X. Zhang and M.J.F. Gales

Research Interests

Kernel methods for speech processing: acoustic modelling and language modelling

Discriminative modelling for speech processing: unstructured discriminative models and structured discriminative models

Generative modelling for speech processing: universal generative acoustic/language models and speaker/environmental adaptation/compensation