Communications of the ACM
Researchers Use Machine Learning to Build Taste Language for Whiskey
Virginia Tech data consultant Chreston Miller (left) and researchers Leah Hamilton (center) and Jacob Lahne will help develop at deep learning tool that analyzes descriptive data.
Credit: Trevor Finney / Virginia Tech
Smoke. Coal dust. Fine leather. Dark berry fruit. Coffee grounds. Understanding the words in published whiskey reviews can be confusing for those considering a $130 bottle of bourbon.
A Virginia Tech research project received a grant from The Institute for Creativity, Arts, and Technology to create a tool that finds a common language in a data set of 6,500 published whiskey reviews of about 50 to 100 words each.
The deep learning tool created by the project could lead to consistent whiskey descriptors and could be used for other research that uses descriptive data. The Virginia Tech team will apply natural language processing to analyze whiskey descriptors.
"We don't know anyone else who has tried to take these reviews . . . and systematically analyze them this way," says Jacob Lahne, a researcher in the Department of Food Science and Technology. "These reviews are in metaphorical, messy, natural language. What we're trying to get to is some shared concept about taste."
From TheRoanokeStar.com
View Full Article
No entries found