I’m passionate about building Artificial Intelligence (AI) with greater ability to collaborate with diverse user populations and individuals. Not every end-user will approach collaboration with artificial agents in the same way. For example when collaboratively preparing a meal, some users may feel less comfortable delegating fire-hazardous work on the stove to an AI collaborator, or some may prefer dictating coordination strategies over extending more freedom to the AI agent. Each individual is unique and often breaks with the most common strategies, and socio-cultural populations may also approach human-AI collaboration with differing behaviors and preferences. Agents trained with “one size fits all” models of human behavior that are deployed across a wide diversity of users will struggle to accommodate this inherently multi-modal user behavior, especially behaviors that are underrepresented or absent from training data. To address this problem, I plan to develop Machine Learning (ML) agents that can more effectively adapt to individual users. I will ground this ML research in Human-Computer Interaction (HCI) studies of how people actually interact and collaborate with AI/ML agents.
I will support these endeavors by continuing my work building greater diversity in the AI/ML research community. This inclusion injects fresh ideas and new perspectives to produce more creative and novel research, while also encouraging more informed and effective application of AI/ML to benefit the diverse populations in which we aim to deploy collaborative agents.
I draw upon my interdisciplinary ML and HCI research experience, which consists of work toward the few-shot scene perception required to interact with diverse users and situations in the visual world [Dornadula 2019 et al.], application of social computing techniques to more accurately evaluate actual end-users’ perceptions of ML model behavior [Zhou 2019 et al.], and ongoing interdisciplinary work towards more individualized human-AI collaboration [Narcomey (ongoing) et al.].
My teaching experience in AI and Computer Vision draws from my academic and research experience alongside my aims toward greater diversity in the AI field.
Check out here for more information about my teaching experience.
Diversity and Indigenous Representation in AI
As a member of the Seminole Nation of Oklahoma, it is important to me to build greater representation of Indigenous people, perspectives, and issues in the AI/ML research community. These representation goals are intertwined with my research goals toward human-AI collaboration that is effective with a wider diversity of users, Indigenous people for example. I’ve given a talk at the AISES (American Indian Science and Engineering Society) 2020 National Conference on this topic, and aim to continue similar work in the future.
Check out here for more information about my work about Indigenous people and AI.
My CV captures my academic, research, teaching, and professional experience, which motivates my interests in human-AI collaboration with diverse users alongside my interests in fostering greater representation of Indigenous people in AI research and industry. View here or download the PDF directly.