FUTURE PLANS

• Enhancements to the decoder
  • Grammar-based decoding

  • N-best search

  • Lattice generation

• Towards a speech-to-text toolkit
  • Acoustic front-end and feature extraction

  • HMM training