One of my dreams, when I first arrived in Rome, was to be able, on a hot summer evening, to walk out to my own lemon tree and pick a still-warm fruit to grace my ice-cold G&T. Sixteen years and four removals later, that tree, bought from a lorry at the side of the road, is still with me and, this wi...
Just as I’ve managed to score a major load of lemons and was looking around for recipes, Jeremy naturally comes up with a brilliant answer.
Recommendation engines are everywhere. They let Netflix suggest shows you might want to watch. They let Spotify build you a personalised playlist of music you will probably like. They turn your smartphone into a source of endless hilarity and mirth. And, of course, there’s IBM’s Watson, recommending all sorts of “interesting” new recipes. As part of his PhD project on machine learning, Jaan Altosaar decided to use a new mathematical technique to build his own recipe recommendation engine.
The technique is similar to the kind of natural language processing that powers predictive text on a phone, and one of the attractions of using food instead of English is that there are only 2000–3000 ingredients to worry about, instead of more than 150,000 words.
The results so far are fun and intriguing, and can only get better.Recommendation engines are everywhere. They let Netflix suggest shows you might want to watch. They let Spotify...
Back in February I had retweeted something interesting from physicist and information theorist Michael Nielsen:
“Augmented cooking with machine intelligence”, with interesting remarks on generating food analogies… https://t.co/UluYk6p8TV
I found the article in it so interesting, there was some brief conversation around it and I thought to recommend it to my then new friend Jeremy Cherfas, whose Eat This Podcast I had just recently started to enjoy. Mostly I thought he would find it as interesting as I, though I hardly expected he’d turn it into a podcast episode. Though I’ve been plowing through back episodes in his catalog, fortunately this morning I ran out of downloaded episodes in the car so I started streaming the most recent one to find a lovely surprise: a podcast produced on a tip I made.
While he surely must have been producing the episode for some time before I started supporting the podcast on Patreon last week, I must say that having an episode made from one of my tips is the best backer thank you I’ve ever received from a crowd funded project.
Needless to say, I obviously found the subject fascinating. In part it did remind me of a section of Herve This’ book The Science of the Oven (eventually I’ll get around to posting a review with more thoughts) and some of his prior research which I was apparently reading on Christmas Day this past year. On page 118 of the text This discusses the classic French sauces of Escoffier’s students Louis Saulnier and Theodore Gringoire  and that a physical chemical analysis of them shows there to be only twenty-three kinds. He continues on:
A system that I introduced during the European Conference on Colloids and Interfaces in 2002  offers a new classification, based on the physical chemical structure of the sauce. In it, G indicates a gas, E an aqueous solution, H a fat in the liquid state, and S a solid. These “phases” can be dispersed (symbol /), mixed (symbol +), superimposed (symbol θ), included (symbol @). Thus, veal stock is a solution, which is designated E. Bound veal stock, composed of starch granules swelled by the water they have absorbed, dispersed in an aqueous solution, is thus described by the formula (E/S)/E.
This goes on to describe in a bit more detail how the scientist-cook could then create a vector space of all combinations of foods from a physical state perspective. A classification system like this could be expanded and bolted on top of the database created by Jaan Altosaar and improved to provide even more actual realistic recipes of the type discussed in the podcast. The combinatorics of the problem are incredibly large, but my guess is that the constraints on the space of possible solutions is brought down incredibly in actual practice. It’s somewhat like the huge numbers of combinations the A, C, T, and Gs in our DNA that could be imagined, yet only an incredibly much smaller subset of that larger set could be found in a living human being.
The additional byproduct of catching this episode was that it finally reminded me why I had thought the name Jaan Altosaar was so familiar to me when I read his article. It turns out I know Jaan and some of his previous work. Sometime back in 2014 I had corresponded with him regarding his fantastic science news site Useful Science which was just then starting. While I was digging up the connection I realized that my old friend Sol Golomb had also referenced Jaan to me via Mark Wilde for some papers he suggested I read.
T. Gringoire and L. Saulnier, Le répertoire de la cuisine. Dupont et Malgat, 1914.
I always prefer a chili recipe with a higher proportion of meat, so this recipe goes much heaver in that department than most. Naturally, high quality ground beef can be substituted for the somewhat healthier turkey if preferred. The beans can be cooked in with the chili simultaneously, but I typically prefer to cook them separately for better doneness and quality as well as well as closer control of the overall soupiness of the chili.
3 tablespoons of olive oil or vegetable oil
6 oz tomato paste
6 tablespoons chili powder
2 tablespoon ground cumin
3/4 – 1.5 teaspoon cayenne pepper (depending on one’s tolerance for heat)
Two 28-ounce cans of (fire-roasted) diced tomatoes
1/4 – 1/2 cup unsweetened cocoa powder
3 tablespoons red wine vinegar
2 tablespoon dried oregano
1 medium to large onion, diced
1 red bell pepper, diced
2 green peppers, diced
2 cups crushed corn tortilla chips
3 cloves garlic, finely minced
4 stalks of celery, finely diced
4 carrots, finely diced
Ingredients for pinto beans
3.5 cups of pinto beans
1/2 onion chopped
1 clove of garlic
1.5 teaspoon of bouillon (or 3 cups of chicken broth)
1/2 teaspoon Kosher salt
Optional ingredients for toppings and serving
sliced scallions or 1/2 raw onion chopped
shredded/grated sharp cheddar cheese
chopped (pickled) jalapeno
corn tortilla chips (or cornbread or white rice)
Bring the pinto beans, onion, garlic, and bouillon in a large pot with an equal amount of salted water to a low boil. Then reduce the heat and cook on low for 3-4 hours until done. Add additional water if necessary during coooking, but don’t allow the beans to become too soupy. Stir regularly to prevent burning to the bottom of the pan.
Put the tomatoes, celery, carrots, onions, peppers, cocoa powder, vinegar, oregano, garlic, crushed tortilla chips, and a teaspoon salt into a covered 6+ quart slow cooker over low heat for 6 hours.
While the above are beginning to cook, heat the oil in a large nonstick pan over medium-high heat with the tomato paste, chili powder, cumin and cayenne and cook for about 2-3 minutes, stirring regularly, until the mixture is dark red and dry in texture. Add the ground turkey, previously seasoned with 1 teaspoon salt, and cook while stirring and breaking up into smaller pieces, until mixture is thoroughly combined. (The turkey doesn’t need to be cooked all the way through but should ideally be browned for better maillard reaction and subsequent flavor).
When the oil, paste, and turkey mixture is done, mix it in with the tomatoes, celery, carrots, et al, and finish cooking. Stir occasionally.
As the turkey/vegetable portion and the beans are done, mix them together in equal measure, and season with salt and pepper to taste.
Serve with scallions, cheddar, sour cream, and pickled jalapeno over tortilla chips. (One could also substitute cornbread or even rice for the tortilla chips for alternate variations.)
Optional cocoa powder may seem a bit out of place in most chilies, but it can serve two functions here: it adds some depth of flavor (without being chocolaty as one may expect) while simultaneously thickening the sauce in the chili.
The celery, carrots, onions, and peppers are all also optional: they can be used to enhance/modify taste, but also add to not only the overall heartiness, but make the dish more veggie friendly for children without detriment to flavor or presentation.
I suggest serving the chili on a bed of tortilla chips (which can also function as a makeshift spoon or eating implement), but it can also be great with cornbread or even served over rice as additional options.
Leftovers can be refrigerated or even frozen (for several weeks) if necessary.
flour (all purpose generally yields better results than cake)
unsalted butter (cold)
fruit: usually dried currants, raisins, chocolate chips, or other fruit
fruit zest (orange, lemon, grapefruit, other)
Other fats could be substituted for the butter, but butter generally tastes best here. For the small handful of health conscious non-professional home cooks, absolutely do not substitute milk for the cream, otherwise the fat ratio for the recipe will be thrown completely off and your results will be horrifying.
5 parts flour : 1 part sugar : 1.5 parts butter : 1 parts egg : 2 parts cream : 1.5 parts fruit
Other ingredients (approximately per part)
1/2 teaspoon salt per part
1 tablespoon baking powder
1/4 oz zest
Professional kitchens scaling the recipe beyond 75 oz of flour, may wish to use 1.25 parts of sugar for more even results.
Preheat oven to 425° F.
Whisk together the flour, sugar, baking powder and salt until mixed thoroughly.
Cut the cold butter into the flour mixture with a pastry blender until the lumps of butter are just larger than the size of a pea. Any smaller and the scones will be tougher and less flaky.
Mix together the cream, egg, (optional currants, raisins, fruit), and the zest, then mix into the flour/butter just until the dough comes together.
Do not overwork the scone dough or the resultant scones will not be light and flaky. You should preferably be able to still see small chunks of butter in the dough.
Roll the dough out into a disk about 1.5″ thick.
Brush a light layer of cream (or milk) onto the top of the disk and sprinkle on a nice layer of cinnamon and sugar.
Using a dough scraper cut the dough into eight equal wedges and place onto cooking sheet.
Put the sheet of scone dough into the oven at 450 for 12-15 minutes until golden brown, or until an inserted toothpick comes out clean.
Cool for a few minutes and then enjoy fresh with clotted cream and fresh fruit.