Mark Gales

Research Interests

• Large vocabulary continuous speech recognition
• Automatic Assessment of Spoken English
• Speaker adaptation and robust speech recognition
• Machine learning (particularly model selection and kernel-based methods)
• Statistical Parametric Speech Synthesis

A brief introduction to speech recognition is available online. For a technical review see The Application of Hidden Markov Models in Speech Recognition. A draft review of model-based noise compensation schemes is also available.

Research Projects

Current projects:

• BABEL (IARPA funded project) - local web-page
• Automatic Oral Communication Skill Evaluation (Cambridge English funded) - local web-page
• The Alta Institute (Cambridge English funded) - local web-page
• Broad Operational Language Translation (DARPA funded project)
• Natural Speech Technology (ESPRC funded project)
• Generative Kernels and Score-Spaces for Classification of Speech (DARPA funded project) - local web-page
• Adaptive Meeting Transcription - PhD Studentship (Toshiba Research Europe Ltd funded)
• HTK Version 3 - HTK V3.4.1 and examples are available.

• Structured Support Vector MAchines for Speech Recognition (Toshiba Research Europe Ltd funded)
• Rapid and Robust Environment Aware Processing (Toshiba Research Europe Ltd funded)
• !CoreTex (Improving Core Speech Recognition technology)
• HTK Rich Audio Transcription (DARPA EARS funded project) - local web-pages
• AGILE (DARPA GALE funded project)

Publications

A full list of publications can be viewed here.

Students

Current list of PhD Research Students and their general research topics:

• Jeff Chen: Adaptive Meeting Transcription systems (Toshiba funded)
• Eric Wang: Universal Acoustic Models (RATS/AGILE/Google Research Award part-funded)
• Chunyang Wu: Canonical Models for Speech Processing (Natural Speech Technology funded)
• Justin Yang: Non-Parametric models for speech recognition (EPSRC funded)
• Austin Zhang: Structured Discriminative Models (Toshiba Funded)

Past PhD Research Students

• Catherine Breslin : Complementary System Generation and Combination [pdf]
• Martin Layton : Kernel Methods for Classifying Variable Length Data [pdf]
• Hank Liao : Uncertainty Decoding for Noise Robust Speech Recognition [pdf]
• Andrew Liu : Discriminative Complexity Control and Linear Projections for LVCSR [pdf]
• Chris Longworth : Speaker Verification and Identification using Kernel Methods [pdf]
• Anton Ragni : Discriminative models for speech (examined)
• CK Raut : Discriminative Adaptive Training and Bayesian Inference for ASR [pdf]
• Antti-Veikko Rosti : Linear Gaussian models for speech recognition [pdf]
• Rogier van Dalen : Noise Robust Speech Recognition [pdf]
• Khe Chai Sim : Structured Precision Matrix Modelling for Speech Recognition [pdf]
• Nathan Smith : Using augmented statistical models and score spaces for classification [pdf]
• Mathew Stuttle : Formant-like features for speech recognition [pdf]
• Kai Yu : Adaptive Training for LVCSR [pdf]

If you are interested in studying for a PhD in the Machine Intelligence Laboratory please see the "Joining Us" page for information.

Teaching

2012-2013 Prof Gales was on Sabbatical this year.

2011-2012

For the Engineering Department:

• MPhil in Advanced Computer Science:
  • Machine Learning for Language Processing (L101)
  • Spoken Language Processing (L106)
• Fourth year engineering undergraduates Module 4F10: Statistical Pattern Processing.
• Material available online (local access only)
• Fourth year engineering undergraduate projects offered 2011-2012 (local access only).

For Emmanuel College: (supervision timetable local access only)

• Part IB Paper 6: Information Engineering

• Large vocabulary continuous speech regonition