Скачать или смотреть Interactive Scalable Visualizations for Data Discoveries and Interpretable AI

Keynote talk for SIGMOD 2021 Curated Session: Interactive Data Exploration

Machine learning and AI models are growing increasingly available, but many models offer predictions that are difficult to understand, evaluate and ultimately act upon. We present how interactive visualization can be used to amplify people’s ability to understand and interact with large-scale data and complex models. We sample from projects where interactive visualization has provided key leaps of insight: Summit and Bluff, systems that scalably summarize and visualize what features a deep learning and how they may be exploited by attacks; SkeletonVis, the first interactive tool that visualizes attacks on human action recognition models; CNN Explainer and GAN Lab (with Google Brain), popular tools for students and experts alike to learn about AI models; Argo Lite, for easy in-browser visualization and sharing of graph visualizations used by thousands of students; CardiacAR, for interactive cardiovascular surgical planning leveraging augmented reality. We conclude by highlighting the next visual analytics research frontiers for data discoveries and AI.

By Prof. Polo Chau https://www.cc.gatech.edu/~dchau/