Improved Context provided by Command Hypotheses Prediction, 2nd iteration

Abstract

Recently, the project AcListant® has achieved command error rates below 1.7 % for automatic speech recognition based on Assistant Based Speech Recognition (ABSR). It was validated that not only significant controller workload reductions were possible, but also significant improvement for ATM efficiency. One main issue to transfer ABSR from the laboratory to the ops-rooms is its costs of deployment. Currently each ABSR model must manually be adapted to the local environment due to e.g. different accents and deviations from standard phraseology. MALORCA proposes a general, cheap and effective solution to automate this re-learning, adaptation and customisation process to new environments, taking advantage of the large amount of speech data available in the ATM world