To address the challenges in generating multimodal instruction data, we developed ProVision, a scalable, programmatic framework that employs scene graphs and human-written programs to systematically synthesize vision-centric instruction data.
author name
Jieyu Zhang
author title Research Intern
My Background
Jieyu Zhang is a Ph.D. student in Computer Science at the University of Washington, Seattle. Before that, they were an undergraduate student in Computer Science at the University of Illinois Urbana-Champaign. My research interests are interactive and data-centric AI/ML with an emphasis on faithful evaluation and effort-light approaches with applications in natural language processing, computer vision, foundation models, science, etc.
My Expertise
Jieyu Zhang's expertise centers on artificial intelligence (AI) and machine learning (ML), with a focus on designing systems that make working with data and AI models more interactive, efficient, and trustworthy. Jieyu Zhang is skilled at creating tools and methods to ensure AI systems are evaluated rigorously, producing reliable results that users can trust. This involves minimizing the effort required to work with complex models and datasets, making AI more accessible and easier to integrate into practical applications.
When I'm Not Writing ...
When I'm not writing, I enjoy playing pool and shooting hoops on the basketball court. These activities help me unwind and keep my energy levels up, offering a great mix of focus and fun. I also love spending time reading, diving into books that expand my knowledge or simply offer a good story to get lost in.
