[Univ of Cambridge] [Dept of Engineering]

Segment Modelling Interest Group


A colloboration of Speech Research Group members interested in various aspects of segment modelling. This colloboration has two primary aims:

  • conduct new research in the area
  • tight integration with state-of-the-art in speech recognition
To achieve these aims, the SMIG will develop and examine:
  • Hidden Markov Model Toolkit (HTK) based segment modelling framework
  • generative and discriminative score-spaces
  • log-linear and other discriminative models


Interested in joining - contact Anton Ragni


  • J. Yang, A. Ragni, M.J.F. Gales, K.M. Knill, Log-linear System Combination Using Structured Support Vector Machines, accepted to Interspeech 2016, pdf