Improve trust in complex genomic data analysis with explainable artificial intelligence

There are many AI applications in every domain, including the medical domain.  However, few AI applications are used in medical domain experts’ daily decisions because most AI computational processes are like black boxes.

In my research, I first get users’ requirements by running an immersive technology survey of medical innovation and implementing two usability studies with domain users. I analysed the usability studies results and addressed the users’ explainable AI needs for the user-centred design.
I then designed and developed interactive and linkable visualisation techniques to explain the complex AI decision process and decision results to clinicians. I created one InTrust model with different XAI methods. My research would make AI more transparent, trustworthy, and ethical for domain users, and then they would eventually use AI to assist their decisions.