Recent advancements in Generative Adversarial Networks (GANs) in creating new data, such as images and videos, have enabled new and interesting trends in deep learning. In this project you are required to investigate and develop a GAN model that can take in speech samples and generate new speech data. This is particularly useful for speech recognition tasks where reaching practicality and high recognition accuracy is unsatisfactory as the results of scarcity of data, for example in impaired automatic speech recognition. You are also required to deploy your model and create a mobile app prototype that inputs an utterance and generate a new one using the deployed GAN model.
· A GAN model that generates new speech data with similar attributes to the input speech signal
· Deployment of the model on the cloud and the development of the required APIs
A mobile app for testing and showcasing the syste
Lab allocations have not been finalised