Investigation on N-Gram Approximated RNNLMs for Recognition of Morphologically Rich Speech.
Balázs TarjánGyörgy SzaszákTibor FegyóPéter MihajlikPublished in: SLSP (2019)
Keyphrases
- n gram
- recognition engine
- language model
- finite state transducers
- speech recognition
- language modelling
- word segmentation
- variable length
- bag of words
- text classification
- language independent
- word recognition
- language modeling
- part of speech
- speech signal
- speech corpus
- finite state
- viterbi algorithm
- word level
- out of vocabulary
- automatic speech recognition
- decision trees
- neural network
- probabilistic model
- hidden markov models
- bayesian networks
- artificial intelligence