Improving EEG based Continuous Speech Recognition
This work addresses incremental improvements in EEG-based speech recognition for assistive technology applications.
The paper tackles improving continuous speech recognition from EEG signals by initializing recurrent layer weights meaningfully and using an external language model, resulting in enhanced performance metrics.
In this paper we introduce various techniques to improve the performance of electroencephalography (EEG) features based continuous speech recognition (CSR) systems. A connectionist temporal classification (CTC) based automatic speech recognition (ASR) system was implemented for performing recognition. We introduce techniques to initialize the weights of the recurrent layers in the encoder of the CTC model with more meaningful weights rather than with random weights and we make use of an external language model to improve the beam search during decoding time. We finally study the problem of predicting articulatory features from EEG features in this paper.