Topic

The TAME Pain study seeks to develop a trustworthy algorithm that can distinguish between different levels of pain by identifying bioacoustic markers in audio data. Through our production of a large dataset complete with extensive annotations, researchers can utilize these resources to explore the connection between vocal patterns and acute pain, advancing the potential for developing novel pain assessment technologies. Introducing this project in a talk pertains to fields of machine learning, healthcare, pain, speech, and audio studies. In addition to its applicability across various disciplines, this study also offers insight to cross cultural collaboration, ethics in protocol development, data collection from human subjects, and data conversion.

The Trust Prism project, currently in its experimental design phase, takes on a broader approach by exploring the most effective presentation of AI recommendations for the greatest alignment of trust. We plan that participants will engage in a task that is supported with an AI recommendation, varying elements such as explanation quality, uncertainty quantification, and difficulty of the task. These variations will enable us to assess how different attributes influence performance, decision-making, and user trust. With our results, we aim to contribute valuable insights to guide the design of AI systems that optimize the performance of human-AI collaboration. 

Presenter

I am an undergraduate student at the University of Texas at Austin studying biochemistry, and my research interests lie in the practicality and responsibility of AI in healthcare. I am really interested in how the Responsible Digital Futures group focuses on sustainable technologies and their responsible deployment, and I hope to explore this focus with the TAME Pain and Trust Prism project.

Connection details
________________________________________________________________________________

Microsoft Teams Need help?

Join the meeting now

Meeting ID: 353 926 500 432 2

Passcode: CV32jr2z