Have you ever eaten a jelly bean while plugging your nose?
(I know… This is probably a question you don’t get all that often.)
I dare you to give it a try. And, better yet, do it without knowing what color jellybean you are eating and then see if you can guess (correctly) which color it is.
I challenged a group of IBMers to just that last week as I led them through a few sensory exercises and a guided tasting of some gorgeous, diverse wines from Domaines Barons de Rothschild (Lafite) vineyards in France, Argentina and Chile.
IBM hired me to lead a discussion on data integrity, data accuracy and wine and I spiced it up with a few sensory activities expressing how critical it is to use data covering the whole picture to get solid results… even in wine tasting and eating.
I wasn’t surprised that the group was not very successful in identifying the jelly bean colors as they ate them with their noses plugged.
Even though it may not appear that jelly beans have much of an aroma, as one chews them, the aromatics are released into the throat, which activates the sense of smell and enables the full flavor to be realized. When the nose is plugged, the airflow is blocked and the aromatics don’t get noticed. And just the sense of taste on the tongue is not enough to distinguish the flavor.
Blocking the sense of smell keeps some of the data away and causes inconclusive results.
It is just like that with Machine Learning and Artificial Intelligence.
For years we’ve all heard “Garbage In, Garbage Out.” As ML and AI grows in popularity, it is often thought that just throwing data into a model will yield great results. I’ve been in countless meetings where I was asked why VineSleuth doesn’t just scrape tasting notes and blog posts or use crowd-sourced data. Aside from the ethics challenge of some of that, there is also a data accuracy and integrity issue at play.
The results of an ML or AI project will only be as good as the data that is used to fuel it. Which is why, at VineSleuth, we are committed to rock-solid analysis of all the wines in our database. We know that’s the only way to get rock solid results… and don’t we all want our results to be true?
Who wants to chew a jelly bean without enjoying its flavor, which comes from the whole picture? Who wants to work with only partial data… especially when solid data is there? Not me.
If you’d like to know how VineSleuth can help your business sell more wine or beer to the right shoppers or diners with AI and sensory science-powered personalization platforms, or if you’d love to just talk AI and solid data in food and beverage, please drop me a note. I’d love to talk with you about either of those.
Founder and CEO, VineSleuth