Craft brewers are running out of beer names. That's what scientist Janelle Shane (who uses artificial intelligence for this purpose frequently) decided to do. THIS sounds like a job for..." In a later article about her neural network-generated metal band names, I told Shane to "Please send me the beer name dataset if you have it." Will artificial intelligence solve the great beer naming crisis?
So what happens when you employ a neural network to create new craft beer names? Researcher Janelle Shane did it by feeding a neural network a bunch of beer names from BeerAdvocate's database -- the project was initiated by Gizmodo's Ryan Mandelbaum -- and the AI came up with a bunch of names which mostly sound very odd, just like real craft beer names. Some examples include Indian Pale Ale names like Dang River, Toe Deal, Earth Pump, Heaven Cat, and Heart Compost, all of which (with possible exception of Toe Deal) I'd devour in a second. Strong Pale Ale and Amber Ale names the AI has conjured are even stranger, with highlights being Slambertangeriss, Brother Panty Tripel, The Vunker the Finger, Thrennt Rem Wine Barrel Aged Monkay Tripel, Gate Rooster, Rickin Organic Red Deaath, and River Smush Hoppy Amber Ale.
I often tell people not to worry about robots taking our jobs, but if you're in the business of branding craft beer, your career might well be nearing its end, thanks to a clever new AI that's figured out how to do it on its own. Researcher and electrical engineer Janelle Shane, who's previously taught AI to conjure up pick-up lines and christen guinea pigs, developed a neural network system to name craft brews by learning from an extensive dataset spanning 90 types of beer culled from BeerAdvocate.com. Here are some of my favorites, which I'd happily consider drinking if I spotted these names on a tap or label: Do those sound good or what? There's more where that came from over at Shane's blog, and you can also hand over your email address for a PDF with 100 more great names.
Now, she's turned her AI naming capabilities towards beer. While writing about her neural network-generated paint color names, Gizmodo's Ryan Mandelbaum mentioned the issues craft brewers were having coming up with names for their beers -- issues that have at times led to legal action. Not long after that, with the help of a Gizmodo reader who put together a dataset of beer names culled from BeerAdvocate.com, Shane plugged the names into a neural network and out popped a bunch of new names for brewers to use. "For these names I turned the neural network's creativity variable higher and got results that can be described mainly as ... interesting," she said on the blog.
For Great Lakes Brewing Company, meeting growing demand means running its processes all day and all night. While Great Lakes had been moving some processes toward automation since Blystone started four years ago, the systems were not connected in a single interface -- and that meant multi-tasking, overworked engineers had to flit from sensor to sensor to check each one, often in the middle of the night. A hackathon with Microsoft and Rockwell came up with the idea to layer analytics and natural language processing on top of the automation process, and in less than two years, Shelby was born. Mike Pantaleano, Rockwell Automation's global business manager, said the firms also have an integration with Microsoft's voice-activated assistant, Cortana, on tap.
First, we should remember that two people may buy the same brand for the same reasons (Miller Lite tastes great). However, they may also buy the same brand for different reasons (Miller Lite tastes great versus Miller Lite is less filling), or buy different brands for the same reasons (Miller Lite tastes great versus Bud Light tastes great) or buy different brands for different reasons (Miller Lite is less filling versus Bud Light tastes great). There are many other ways qualitative can help inform quantitative research, but a fuller elaboration would require much more space than I have here. And, of course, we all can be guilty of condemning sweeping generalizations with... sweeping generalizations!
That's where a big rig tricked out with a sophisticated system that lets a computer take control on the road delivered 50,000 cans of Budweiser last week -- in what the beer company says was the first commercial delivery using the tech. Unlike other self-driving systems on the market, such as Tesla's autopilot, Otto's tech lets drivers get out from behind the wheel altogether. Industry watchers have long predicted that autonomous vehicles would transform businesses like trucking and mining before taking over the consumer market. That could mean big business for companies like Otto because nearly 70 percent of all freight in the U.S. is delivered via trucks, according to American Trucking Associations.
Enter Plum, Koretz's 1,499 invention that doubles as a countertop appliance and the ultimate tricked-out wine storage. Plum lets wine drinkers have a "by the glass" experience: When consumers place a bottle in the device, Plum uses cloud and artificial intelligence technology to identify the label, and automatically set the perfect sitting temperature for each bottle. Koretz said the company has raised nearly 10 million to date from lead investor Vinod Khosla, of Khosla Ventures, as well as investors across various industries - wineries, consumer goods, hospitality and technology. "We want to make Plum an extension of the tasting room," he said.
The machine learning algorithms uses a combination of reinforcement learning and bayesian optimisation to assist the brewer in deciding how to change the recipe of the beer, with the algorithms learning from experience and customer feedback. The machine learning approach used to brew the new beers is a Bayesian technique, in the form of Bayesian non-parametrics. In the same way augmented reality technologies will sit between our own vision of the world, mixing the simulated and the real world, machine intelligence will sit between our actions and the world. "The brewer's interaction with the algorithm is very much part of the technology and generates additional learning data--we like to think of the algorithm as being a bit like an apprentice, listening and learning from a master," says McInerney, "However, unlike a human apprentice the tech can listen to multiple different master brewers as well as all the customers at the same time."
To create the perfect batch, the London-based brewing company behind the project collects customer feedback through a Facebook Messenger bot then inputs that information into an algorithm to tweak the beer's recipe. IntelligentX affectionately calls its algorithm ABI for Automatic Brewing Intelligence. IntelligentX beers are brewed in four varieties -- golden, amber, pale and black -- though it is unclear how curious drinkers can get their hands on a bottle. A cryptic message on the company's website reads: "Our IntelligentX beer comes in exclusive'pirate black' bottles and is available in extremely limited editions for those that know how to find it…" Those who are interested can sign up with an email address at IntelligentX.ai.