As part of her PhD, Ashleigh developed a cloud-based web application to enable non-technical users to create and train their own Neural Machine Translation (NMT) network, either from scratch (with full control over the design of a transformer or recurrent neural network based encoder decoder network) or fine-tuned from existing state-of-the-art networks.
The tool was intended to broaden the demographics of researchers able to easily conduct high-quality, replicable NMT research, and to demonstrate a proof-of-concept of a framework for research tool design.
The development of this tool followed a user-centered design process, to ensure that non-technical users were able to operate this tool successfully.