Advancing human-AI collaboration: multimodal systems for accessibility, research, and learning
OA Version
Citation
Abstract
This thesis explores domain-specific AI applications in assistive technology, education, and botany. First, CameraMouse, an AI-powered system, enables computer use for individuals with severe motor impairments by converting head movements into cursor control. User studies highlight its effectiveness in enhancing accessibility. Next, an open-source platform for developing AI-powered academic chatbots was developed. Lastly, the thesis addresses species recognition and analysis in herbarium research, using fine-grained image classification, zero-shot learning, and conversational models. Early experiments show promising results in identifying subtle species differences. These projects demonstrate AI’s impact on accessibility, education, and botany, and are unified by a common goal: designing intelligent, human-centered systems that empower users and make workflows more effective.
Description
2025
License
Attribution 4.0 International