Google Employees Try Baking Recipes Created by AI (foodandwine.com) 27
"Behold the cakie: It has the crispiness of a cookie and the, well, 'cakiness' of a cake."
So says a triumphant blog post by Google Cloud's developer advocate and an applied AI Engineer for Google's Cloud AI. "We also made breakies, which were more like fluffy cookies, almost the consistency of a muffin" (or bread).
Food and Wine explains the project (in an article shared by Slashdot reader John Trumpian): Inspired by the pandemic-spawned spike in searches for baking, the team at Google Cloud "decided to dive a little deeper into the trend and try to understand the science behind what makes cookies crunchy, cake spongy and bread fluffy," according to a post on their blog. Then, once armed with that machine learning knowledge, they attempted to mix these attributes into what they bill as "two completely new baking recipes...."
[T]hese Google Cloud employees organized about 700 recipes covering cookies, cakes, and breads — standardizing measurements, isolating the key ingredients, and re-categorizing things like banana breads that aren't really "breads." Then, they fed them into a tool called "AutoML Tables" to create a machine learning model that was able to predict whether a recipe was a cookie, cake, or bread based on its ingredient amounts. ["If you've never tried AutoML Tables, it's a code-free way to build models from the type of data you'd find in a spreadsheet like numbers and categories — no data science background required," explains the blog post.]
Of course, recipes don't necessarily fit perfectly into one category. As Sara Robinson, who led the project, explained, a recipe might come back as 97 percent bread, 2 percent cake, and 1 percent cookie. So what if she asked the model to create its own recipe: something that's 50 percent cookie and 50 percent cake?
That's how the Cakie was born. And she was happily surprised by the results. "It is yummy," Robinson said. "And it strangely tastes like what I'd imagine would happen if I told a machine to make a cake cookie hybrid." Based on that success, she and colleague Dale Markowitz continued to tweak their model — which resulted in the Breakie.
"We should caveat that while our model gave us ingredients, it didn't spit out any baking directions, so we had to improvise those ourselves," the blog post explains. "And, we added chocolate chips and cinnamon for good measure." Robinson also built a prediction-making web app to help quickly experiment with different ingredient ratios.
They ultimately identified which ingredients were the biggest "signal" of cake-ness, cookie-ness, and bread-ness, concluding that "In our case, the amount of butter, sugar, yeast and egg in a recipe all seemed to be important indicators..."
So says a triumphant blog post by Google Cloud's developer advocate and an applied AI Engineer for Google's Cloud AI. "We also made breakies, which were more like fluffy cookies, almost the consistency of a muffin" (or bread).
Food and Wine explains the project (in an article shared by Slashdot reader John Trumpian): Inspired by the pandemic-spawned spike in searches for baking, the team at Google Cloud "decided to dive a little deeper into the trend and try to understand the science behind what makes cookies crunchy, cake spongy and bread fluffy," according to a post on their blog. Then, once armed with that machine learning knowledge, they attempted to mix these attributes into what they bill as "two completely new baking recipes...."
[T]hese Google Cloud employees organized about 700 recipes covering cookies, cakes, and breads — standardizing measurements, isolating the key ingredients, and re-categorizing things like banana breads that aren't really "breads." Then, they fed them into a tool called "AutoML Tables" to create a machine learning model that was able to predict whether a recipe was a cookie, cake, or bread based on its ingredient amounts. ["If you've never tried AutoML Tables, it's a code-free way to build models from the type of data you'd find in a spreadsheet like numbers and categories — no data science background required," explains the blog post.]
Of course, recipes don't necessarily fit perfectly into one category. As Sara Robinson, who led the project, explained, a recipe might come back as 97 percent bread, 2 percent cake, and 1 percent cookie. So what if she asked the model to create its own recipe: something that's 50 percent cookie and 50 percent cake?
That's how the Cakie was born. And she was happily surprised by the results. "It is yummy," Robinson said. "And it strangely tastes like what I'd imagine would happen if I told a machine to make a cake cookie hybrid." Based on that success, she and colleague Dale Markowitz continued to tweak their model — which resulted in the Breakie.
"We should caveat that while our model gave us ingredients, it didn't spit out any baking directions, so we had to improvise those ourselves," the blog post explains. "And, we added chocolate chips and cinnamon for good measure." Robinson also built a prediction-making web app to help quickly experiment with different ingredient ratios.
They ultimately identified which ingredients were the biggest "signal" of cake-ness, cookie-ness, and bread-ness, concluding that "In our case, the amount of butter, sugar, yeast and egg in a recipe all seemed to be important indicators..."
UGH! Put this in idle! (Score:1)
Oh well, there's no account for taste
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
yum.
but did this a i consider t h c recipes
Cakie is nothing new (Score:4, Interesting)
Japanese melonpan fits that description to a t.
Interesting for diets (Score:4, Insightful)
That might be interesting for diets. My last girlfriend had a low FODMAP diet (meaning avoid vegetables and fruits high in a bunch of sugars that basically bloat and/or give you gas).
It would be very interesting to let these ML tables generate recipes with the limited stuff people with specific diets can eat. It would be tasty and probably more nutritious as well, because we tended to eat pretty simple ("safe") meals while we were together.
Re: (Score:2)
"It would be very interesting to let these ML tables generate recipes with the limited stuff people with specific diets can eat. "
IBM's Chef Watson is/was great for this, give him ingredient and he spits out recipes.
But it may have been canceled lately.
Wow (Score:5, Insightful)
These Google employees have just demonstrated why normal people sometimes like to make fun of techies.
Maybe (Score:5, Interesting)
One thing they have demonstrated is baking is a specialized form of chemistry and like chemistry, a machine can be used to yield the desired results. Yes, this was minor experiment but it demonstrates the potential of machine learning. Given more attention, this could be the preface to substituting ingredients in order to meet dietary requirements while the desired food seems the same to the person eating it.
I'm always of fan of interdisciplinary research because sometimes it yields interesting results and new insights.
Re: (Score:3)
Uhh, the usual goal of AI of this sort is to e.g. look for novel new drugs out of millions and millions of combinations/formulations.
In this case, it basically did something many bakers could do by themselves with 30 seconds of thought. Tell a human baker what you're looking for and they would be able to tell you the same recipes. Pretending "ooh, how novel" when baking (at this level) is well understood is a bit silly.
Now, if it had come up with some totally new texture and a flavor no human had done befor
Re: (Score:2)
Just because an AI can now do the same thing that a human can do doesn't mean that it isn't useful. Anyone can read a map and find out how to navigate with minimal instruction, but turn by turn navigation sure made that process easier for a billion people.
Cooking itself is a fairly easy skill for instance. A lot of people struggle with the ingredients side; what can I make with the stuff in my pantry or at a local grocery and how to do it. If pattern matching algorithms can identify reliable substitutions f
Re: (Score:2)
Re: (Score:2)
Especially if those "normal people" have any knowledge of food & baking... Because nothing they describe sounds new at all. Hell, I was just watching America's Test Kitchen bake a hybrid cake/cookie earlier this week.
It can never go beyond what it's trained for. (Score:2)
It is not an AI. It is a pattern fuzzer.
One with really shit pre-assumptions based on what a bunch of clueless wannabe chefs know about cooking.
Which is as close to clueful as when some Hollywood writers try to write a "hacker" character.
Re: (Score:2)
Fart around with categories then get baked (Score:1)
The use of AI (Score:2)
Seems the use of AI wasn't really used in making receipts, but to help them understand how to bake and which role the ingredients play.
It doesn't take much to understand baking ingredients however. Sugar crystallise under a lot of heat and so will add a crunch if one bakes enough sugar into it. Eggs don't, but remain soft when cooked, baked or fried and therefore help in keeping a cake spongy. Yeast helps in making a product fluffy, because it produces gas, which is why one lets the dough rise for a while t
Douglas Adams had this with Deep Thought (Score:4, Funny)
What an advance (Score:3)
AI ? (Score:1)
Congratulations! (Score:1)
GLADos (Score:1)
This was a triumph! I'm making a note here...HUGE success!
Anyway, this cake is great--so delicious and moist!