Empowering Community Enterprises with AI

Overview

We focus on developing cutting-edge AI solutions that deliver tangible impact across diverse enterprise contexts—especially those committed to social good, equity, and community empowerment. Our team brings deep expertise in vision technologies and excels in achieving high-performance outcomes even with limited data—an essential capability when working with equity-focused groups and underrepresented communities. 

Members of our team work on projects that exemplify this mission by tackling a critical accessibility challenge through the development of robust Auslan (Australian Sign Language) translation technologies. Funded by the Australian Research Council (ARC) and Google, and delivered in partnership with DeafConnect, this initiative is deeply informed by the lived experiences and needs of the Australian Deaf community. It demonstrates how enterprise AI can be purposefully applied to remove communication barriers, such as those faced by Deaf employees in workplaces, thereby expanding employment opportunities and fostering inclusive enterprise environments.

To support this, we curated MM-WLAuslan, the first large-scale, multi-view Auslan translation and recognition dataset. This dataset enables models to perform reliably across varied camera angles—a key requirement for real-world deployment. Given Auslan’s smaller user base and limited data availability compared to languages like ASL, we applied advanced techniques such as transfer learning and feature disentanglement to leverage data from other sign languages. These innovations significantly boost model accuracy and robustness, showcasing our team’s ability to deliver enterprise-grade AI solutions that are both inclusive and high-performing.

Visit: Auslan-daily Dataset