How should VIS4ML Redefine Itself in the Rapid Evolution of AI?
Date&Time to be determined
IEEE VIS 2023, Melbourne, Australia, Oct. 22-27 🔗
We propose a panel to discuss the changing role of visualization in the development and deployment of machine learning models in light of the rapid evolution of artificial intelligence (AI). Visualization for machine learning (VIS4ML) has been a thriving research area within the visualization community because of the need for better affordances and representations to enable broad groups of stakeholders to interact with and interpret machine learning models. However, recent advancements in AI are changing our understanding of the capabilities of machine learning models, both in performance and in their ability to interact with the general population. In light of these advancements, we feel it is an important time for the visualization community to consider how the opportunities for visualization have changed. We have gathered a diverse set of panelists from both academia and industry, with varying levels of experience. We hope that providing a multitude of perspectives will shed light on new opportunities for visualization research, while providing context on the natural evolution of the field over the last few decades. The panel format will begin with introductory statements from each panelist. Then, through a set of open-ended questions, we will ask panelists to have an open discussion about which types of stakeholders, use cases, and steps of the modeling pipeline they expect to change the most. The panel will conclude by asking each panelist to share where they feel the best opportunities are for VIS4ML research in the medium term future.