Advance Queensland has recently funded a new Trusted Autonomous Systems Project led by academics within the School of Information Technology and Electrical Engineering (ITEE).
The project will build Queensland's capacity in AI language technologies and secure confidential, on-shore speech technologies which are essential for data sovereignty in cultural, health and security services.
Current commercial practices process speech data off-shore and international technology companies routinely use clients’ data in their speech training algorithms. Research is urgently needed in the area of secure, confidential, on-shore processing of speech and language data to make it more practical, widely available and integrated within the digital asset managers that underpin AI language technologies.
Pama Language Centre (PLC) and Professor Janet Wiles and Dr Ben Matthews - alongside Ben Foley from the School of Languages and Cultures - will collaborate on a series of projects with speech communities. They will be designing, developing and evaluating new language technologies including a digital asset manager for language resources, tools to support sharing of augmented reality assets, and workshops to build capacity and resource creation.
It will also extend the ARC Centre of Excellence for the Dynamics of Language (CoEDL) Transcription Acceleration Project (TAP) application of ‘transfer learning’ for speech recognition, aimed at reducing the training data and time needed to develop novel speech recognition systems. An example of an existing PLC project utilising Augmented Reality is available here.
The success of this project was also recently blogged by the Trusted Autonomous Systems team here.
