Why whisky and AI are a perfect cocktail
Introduction
I love whisky. I know, that's kind of a given -- you're reading this blog, after all. But I also love AI and machine learning, so when the two topics come together in discussions about AI-assisted whisky distilling and blending, my nerd senses start tingling. As it turns out, it's not just me: The New York Times has written about this before, too. In fact, with some analysts predicting that AI will be used in every industry by 2030 (including the food industry), there are many reasons why this is such an exciting time for whisky lovers like us:
Over the last few years, there's been a lot of buzz about AI and whisky.
AI is being used to blend different whiskies together, and it could even be used to create entirely new flavor combinations.
The machines may not make the best whisky, but they may help us discover some great new ones.
Whisky startups aren’t limited by tradition. They can use technology to try out all sorts of weird ideas that might not have been tried before if they weren’t so expensive to make and then turn into money-losing experiments.
It started with The Glenlivet, which developed an AI model to help with cask selection.
In 2016, The Glenlivet became the first distillery to use artificial intelligence (AI) in cask selection.
The data scientists at IBM used a model they built called Watson Discovery Advisor to help with this process. The resulting whisky was called “Project Alpha” and it received rave reviews from industry experts.
Diageo tried using AI to evaluate casks.
Diageo, the world's largest producer of Scotch whisky, is using AI to evaluate casks.
A whiskey distiller needs to know which casks will produce the tastiest whiskey. A traditional way of determining this is through a tasting panel, where a group of experts taste each cask and declare whether it's good or bad. However, this method is time-consuming and expensive—and can lead to biased results if some tasters are more experienced than others.
Diageo wanted to find a better way on their own terms: they wanted an algorithm that could predict whether a particular cask would make good whiskey without relying on human input at all (or perhaps in combination with human input).
There have been few attempts at distilling whisky using AI so far.
If you're looking to experiment with AI-based whisky distillation, there's a lot of research being done on the subject. There have been few attempts at distilling whisky using AI so far: in June, the Royal Society of Chemistry published an article detailing how deep learning can be applied to predict the chemical composition of whiskies based on their taste profiles. But if you're curious about how AI can be used in your own home lab, we'd suggest checking out another paper published earlier this year by researchers from Scotland's Heriot-Watt University—it details how they used machine learning algorithms to train a neural network that could classify different types of whiskies by flavor profile and chemical composition.
AI has been used to blend different whiskies together.
While AI can be used to create new flavor combinations or blend different whiskies together, it can also help us develop the next generation of single malt Scotch whisky.
In a recent study published in the Journal of Food Science, researchers from the Heriot-Watt University in Edinburgh analyzed how AI techniques could be used to create a new taste for Scotch whisky. They found that an artificial intelligence model had identified between 50 and 200 flavor compounds found in single malt scotch that had never been reported before.
AI can be used to create many new flavor combinations.
We can use AI to create new flavor combinations that are impossible to achieve using traditional methods. In 2016, a team of researchers from the University of California, Berkley developed an AI algorithm that can predict how new flavors will taste based on their chemical structure. The algorithm has been dubbed NetFlavor and is designed specifically for food applications. By combining this tool with expert knowledge about food chemistry and sensory analysis, we’re able to predict how a particular combination of ingredients will taste before it’s ever made.
In addition to helping you create new flavors, NetFlavor also helps speed up the process of experimenting with different flavor pairings by eliminating some guesswork involved in creating recipes. With this technology available now and more on the horizon (I won't speculate on what else!), you don't have to settle for boring ol' Scotch anymore!
The machines may not make the best whisky, but they may help us discover some great new ones.
The machines may not make the best whisky, but they may help us discover some great new ones. AI can help us find new flavor combinations, production methods and ways to blend whiskies. These are all things that are simply too difficult for humans to do on their own.
This is especially true when it comes to flavors. Flavors are incredibly complex, and there are thousands of different molecules in a given batch of Scotch whisky alone. Even if we were able to identify every single molecule that contributes to its taste—a nearly impossible task—there'd still be an infinite number of ways we could combine them into new recipes. As a result, many companies use machine learning algorithms like neural networks or evolutionary programming techniques that can learn from past data sets (i.e., existing recipes) rather than requiring an expert programmer's hand-crafted rules about how each component should behave in relation to every other component [4].
There's also a lot more room for innovation here than people realize: when whisky makers started using barrels made from exotic woods like oak or sherry cask wood instead of just traditional American white oak barrels (which have been used since the 1700s), they got some surprising results! The flavors imparted by these unusual woods were not what anyone would have expected them to be beforehand; but now that this has happened once before with something else entirely—charcuterie meats cured inside wine barrels instead
Whisky startups aren't limited by tradition. They can experiment with new flavors or new production methods -- and use AI to do it.
The first thing to know about AI is that it's not just a buzzword. It can be used to create many new flavor combinations, blend different whiskies together, and even create a completely new flavor.
In fact, some of the world's leading whisky companies are already using AI in their production processes:
The Glenmorangie Co., which produces more than 3 million bottles of whisky each year at its distillery on the banks of Scotland's Dornoch Firth, uses an artificial intelligence-powered system called "Dermot" to analyze its whiskies' chemical composition as they're distilled. The information helps them improve their craftsmanship and develop new flavors (like a recent collaboration with James Bond star Roger Moore).
Beam Suntory Inc., which owns Jim Beam and Maker’s Mark among other brands, has developed an app called Mixology that lets drinkers search for cocktails by taste profile or ingredient type -- then gives them recommendations based on those preferences.
Wild Turkey Distillery Inc., based in Lawrenceburg Kentucky (a town known as "bourbon capital"), is incorporating machine learning into its bottling process so its automated labeling equipment can identify whether bottles need special care before filling them with liquid gold (and if they're too warm).
In just a few years, we could be drinking whisky that was produced by algorithms that were trained by artificial intelligence systems -- but we won't know it until the bottle is opened.
You might be familiar with artificial intelligence (AI) as the technology behind self-driving cars and voice assistants, but it has also been used in other industries to create new products. In the whisky industry, AI is being used to create new blends and flavors that were previously impossible to re-create by hand.
A few years ago, Glen Moray released a limited edition whisky called The Last Drop using an algorithm called NEAT (Neural Evolution through Augmented Topography) that generated 100 different variations of the same blend, each one slightly different from the last. Each version was then tested by master distillers in order to find out which blend worked best with all its ingredients.
In fact, it wasn't long before a group of researchers came up with another algorithm called Whisky Bot -- this one takes data about a particular type of whisky and uses it as input into a neural network that generates an infinite number of new recipes based on how users interact with its interface (in this case via Twitter). These recipes then get rated based on their characteristics -- lengthiness is ranked from 1-5 stars -- until there's only one left standing at any given time!
This technology isn't just limited to creating new blends either: It can also be used for finding ways around environmental constraints when making whisky locally near where it would be consumed abroad because climate conditions aren't always ideal closer towards home base where most distilleries are located right now."
Machines might soon be able to produce a perfectly consistent whisky for every taste, but will humans want it?
This is not to say that machines will never be able to perfectly replicate the taste of a certain whisky. But at this point, humans are still necessary for the final step: tasting and blending.
In fact, Glenlivet's master blender Richard Paterson has been experimenting with machine learning in his work for years. He was one of the first people from Scotch whisky industry to take an interest in AI when he partnered with IBM to build a computer program called Chef Watson (which can create new recipes). He has also tried using AI as a way of finding new flavors by testing different combinations of ingredients in his lab.
But even if machines can produce consistent products for every person on earth (and maybe someday they will), will humans want it? Will we become content with our robotic bartenders—or would we miss the feeling of being served by someone else?
Conclusion
Whisky is a great place for AI to start. It's a mature industry with centuries of tradition, but it's also one that could benefit from new technologies. AI can help distillers make better whisky and find those perfect barrels in their warehouses, but the machines aren't going to be able to match human creativity any time soon. Distillers are still needed -- they just need some help from their friends at home on the computer screen! This blog post is entirely produced by Artificial Intelligence.
Comentarios