Ebbinghaus Empire: Perceptual semantics for visual communication

Visual communication is fundamental to how humans share information, from weather patterns to disease prevalence, to their latest scientific discoveries. When people attempt to interpret information visualizations, such as graphs, maps, diagrams, and signage, they are faced with the task of ascribing meaning to perceptual features—perceptual semantics. Sometimes, visualization designs include legends, labels, or captions to help determine perceptual semantics in the context of the visualization. However, people have expectations about how perceptual features will map to concepts called inferred mappings, and they find it more difficult to interpret visualizations that violate those expectations. Traditionally, studies on inferred mappings distinguished factors relevant for visualizations of categorical vs. continuous information. In this talk, I will discuss recent work that unites these two domains within a single framework of assignment inference. Assignment inference is the process by which people infer mappings between perceptual features and concepts represented in encoding systems. I will begin by presenting evidence that observers infer globally optimal assignments by maximizing the “merit,” or “goodness,” of assignments between perceptual features and concepts, with an emphasis on color semantics. I will then discuss factors that contribute to merit in assignment inference and explain how we can model the combination of multiple (sometimes competing) sources of merit to predict human judgments. This work has increased our understanding of how people ascribe meaning to perceptual features, which can be used to make visual communication more effective and efficient.

Alternate locations:

Online: https://utoronto.zoom.us/j/88523051719