Many speech-enabled systems in the market have restricted speech capabilities. Systems which can interact with users, usually have limited vocabulary and support only one language at a time. Furthermore, these systems do not have learning capabilities and, as such, their interactive responses are limited.
In this project, the NeuraBASE network model was used as an alternate approach to a human speech recognition system with interactive and adaptive learning capabilities. The mimic feature of this prototype matches the speech input from the user with words previously spoken and found in NeuraBASE. This allows the system to reiterate the input speech. The interactive feature allows the system to respond to a speech input with another phrase. The response to the input speech is based on continuous training by a user.