Martin Layton

Martin Layton is a PhD student in the Machine Intelligence Laboratory (formerly the Speech Vision and Robotics (SVR) group) under the supervision of Dr Mark Gales.

Research interests

• Augmented Statistical Models
• Acoustic Modelling for speech recognition
• Generative kernels (a more general form of Fisher kernel)
• Maximum margin (for augmented models) and conditional maximum likelihood (for conditional augmented models) estimation

Publications

Journal Papers

• M.J.F. Gales and M.I. Layton (2006).
  Training Augmented Models using SVMs,
  IEICE

Conference and Workshop Papers

• M.I. Layton and M.J.F. Gales (2006).
  Augmented Statistical Models for Speech Recognition.
  ICASSP 2006 (to appear)
• M.I. Layton and M.J.F. Gales (2005).
  Augmented Statistical Models: Exploiting Generative Models in Discriminative Classifiers.
  NIPS Workshop 2005
• M.I. Layton and M.J.F. Gales (2005).
  Acoustic Modelling using Continuous Rational Kernels.
  MLSP 2005
• M.J.F. Gales and M.I. Layton (2004).
  SVMs, Score-spaces and Maximum Margin Statistical Models.
  Beyond HMM Workshop 2004, ATR

Technical Reports

• M.I. Layton (2004).
  Maximum Margin Training of Generative Kernels.
  First Year Report, August 2004
• M.I. Layton and M.J.F. Gales (2004).
  Maximum Margin Training of Generative Kernels.
  Technical Report CUED/F-INFENG/TR.484, June 2004

Contact Information