ORLANDO, Aug. 12, 2022 – A UCF readability researcher labored with an Adobe group on a machine studying model to present personalized font suggestions that enhance the accessibility of digital info and improve particular person studying experiences.
The group was comprised of Adobe machine studying engineers and researchers who collaborated with imaginative and prescient scientists, typographers, knowledge scientists, and a UCF readability researcher to research Adobe’s machine studying model generally known as FontMART.
The outcomes have been lately printed in ACM Designing Interactive Systems 2022.
Adobe is a part of The Readability Consortium that leads UCF’s digital readability analysis utilizing individuated typography to improve digital readability for readers of all ages and skills. Adobe’s FontMART analysis was performed in collaboration with UCF’s Virtual Readability Lab.
“The way forward for readability is a tool watching people learn and utilizing their efficiency to tailor the format in order that they learn at their greatest,” says Ben Sawyer ’14MS ’15PhD, the director of the Readability Consortium and UCF’s Virtual Readability Lab. “We look ahead to the day when you may decide up a tool, learn and obtain info in a means that uniquely fits your wants.”
Sawyer and Zoya Bylinskii, an Adobe analysis scientist, have been concerned within the conception of the analysis and supplied steering all through the research. Tianyuan Cai, an Acrobat.com machine studying engineer, led the FontMART research.
The research used the Font Preference Test featured on UCF’s Virtual Readability Lab’s web site to present baselines for evaluating FontMART’s suggestions.
The consideration of font desire is necessary since folks’s most popular fonts typically differ from the font that may greatest enhance their studying expertise and efficiency. The discrepancy between a reader’s most popular font and quickest font has been demonstrated in earlier readability analysis.
Study outcomes indicated that the FontMART model can advocate fonts that enhance studying pace by matching reader traits with particular font traits.
How the Model Works
The FontMART model learns to affiliate fonts with particular reader traits. FontMART was educated with a distant readability research of 252 crowd employees and their self-reported demographic info. Interviews with typographers influenced the collection of the eight fonts used within the research. The remaining font choice included fonts from each the serif (i.e., Georgia, Merriweather, Times, and Source Serif Pro) and Sans Serif (i.e., Arial, Open Sans, Poppins, and Roboto) households.
The impact of a font varies by readers, researchers discovered.
FontMART can predict the fonts that work properly for particular readers by understanding the connection between font traits and reader traits like font familiarity, self-reported studying pace and age, in accordance to the FontMART research. Among the traits thought-about, age performs the biggest position when the model determines which font is really useful for readers.
For occasion, font traits like heavier weight profit the studying expertise of older adults as a result of thicker font strokes are simpler to learn for these with weaker and variable eyesight.
More analysis is required and should embody broader age distribution of individuals to be extra consultant of the overall inhabitants, evaluating the model’s effectiveness for different studying contexts like long-form or glanceable, and increasing the languages and related font traits to higher accommodate reader variety.
Continued collaborations and analysis will assist develop the traits explored to enhance the FontMART model and improve particular person studying experiences.
UCF’s Readability Consortium and Virtual Readability Lab tackle how personalization can enhance studying effectivity and pace. Sawyer additionally leads LabX, an utilized neuroscience group centered on human efficiency, and he’s an affiliate professor in industrial engineering and administration programs. Sawyer acquired a doctorate in human elements psychology and a grasp’s diploma in industrial engineering from UCF. He accomplished his postdoctoral research at MIT.
Method of Research
Experimental research
Subject of Research
People
Article Title
Personalized Font Recommendations: Combining ML and Typographic Guidelines to Optimize Readability
Article Publication Date
13-Jun-2022
Disclaimer: AAAS and EurekAlert! should not accountable for the accuracy of reports releases posted to EurekAlert! by contributing establishments or for the usage of any info via the EurekAlert system.
https://www.eurekalert.org/news-releases/961726
