Kwan-Liu Ma: AI-Assisted and Human-Centered Approaches to Visualization Design

Описание: Keynote by Kwan-Liu Ma from UC Davis at the Visual Computing Trends 2025 symposium in Vienna on January 30, 2025: "AI-Assisted and Human-Centered Approaches to Visualization Design".

The rapid advancement of AI technology is transforming every field, including the domain of visualization. As a powerful tool for data-driven problem-solving, decision-making, and storytelling, visualization must be thoughtfully designed to ensure both effectiveness and performance. I will explore how AI and machine learning can enhance the visualization process, drawing from projects conducted by my research group. These technologies hold significant promise in automating aspects of data analysis, guiding in the visualization design space, and optimizing visual outputs to improve user experience. However, not all facets of visualization can be fully automated. Many visualization tasks require a nuanced, human-centered design approach that involves careful consideration of user needs, context, and iterative evaluation—efforts that AI, at least in its current form, cannot entirely replace. I will discuss the complexities of designing such visualizations, particularly for exploratory purposes, where the goal is to empower users to investigate data dynamically and intuitively. Using case studies focused on creating exploratory visualizations for the general public, I will share key insights and lessons learned, highlighting the importance of balancing technical innovation with thoughtful, user-focused design.