可视化与人工智能交叉研究综述

A survey on interdisciplinary research of visualization and artificial intelligence

随着人工智能技术的突破性进展,人工智能与可视化的交叉研究成为当前的研究热点之一,为人工智能和大数据分析领域的若干核心难题提供了启发式的理论、方法和技术.一方面,人工智能技术的创新应用提升了可视化的分析效率,拓展了分析功能,为大数据可视分析提供了强有力的工具.另一方面,可视化技术增强了以深度学习为代表的人工智能的可解释性和交互性,为可解释人工智能提供了可靠的技术基础.本文从面向人工智能的可视化技术和人工智能驱动的可视化技术两个方向,分别介绍了数据质量改善、可解释机器学习、智能特征提取、可视化自动布局与生成、智能交互、智能故事叙述等6个重要问题.对国内外的研究进展进行了分析,并对发展趋势进行了展望.

With the breakthroughs in artificial intelligence(AI)technology,interdisciplinary research across AI and visualization(AI+VIS)has become one of the current research hotspots,providing inspiring theories,methods,and techniques for several core challenges in AI and big data analytics.On one hand,the innovative application of artificial intelligence technology has improved the efficiency of visualization,expanded the analysis capabilities,and provided powerful tools for big data visualization and analysis.On the other hand,visualization techniques enhance the explainability and interactivity of AI represented by deep learning,providing a reliable technical foundation for explainable AI.This paper introduces six important topics,including data quality improvement,explainable machine learning,intelligent feature extraction,automatic visualization layout and generation,intelligent interaction,and intelligent storytelling from two directions of"VIS for AI"and"AI for VIS",respectively.The research progress in the recent three years is analyzed.We also highlight the research trends of AI+VIS.