So if you’ve ever picked out paint, you know that every infinitesimally different shade of blue, beige, and gray has its own descriptive, attractive name. Tuscan sunrise, blushing pear, Tradewind, etc… There are in fact people who invent these names for a living. But given that the human eye can see millions of distinct colors, sooner or later we’re going to run out of good names. Can AI help?
For this experiment, I gave the neural network a list of about 7,700 Sherwin-Williams paint colors along with their RGB values. (RGB = red, green, and blue color values) Could the neural network learn to invent new paint colors and give them attractive names?
One way I have of checking on the neural network’s progress during training is to ask it to produce some output using the lowest-creativity setting. Then the neural network plays it safe, and we can get an idea of what it has learned for sure.
By the first checkpoint, the neural network has learned to produce valid RGB values – these are colors, all right, and you could technically paint your walls with them. It’s a little farther behind the curve on the names, although it does seem to be attempting a combination of the colors brown, blue, and gray.
By the second checkpoint, the neural network can properly spell green and gray. It doesn’t seem to actually know what color they are, however.
Let’s check in with what the more-creative setting is producing.
…oh, okay.
Later in the training process, the neural network is about as well-trained as it’s going to be (perhaps with different parameters, it could have done a bit better – a lot of neural network training involves choosing the right training parameters). By this point, it’s able to figure out some of the basic colors, like white, red, and grey:
Although not reliably.
In fact, looking at the neural network’s output as a whole, it is evident that:
The neural network really likes brown, beige, and grey.
The neural network has really really bad ideas for paint names.
Interestingly, part of the reason that, until now, Trump has been supported in Russia is due to translation. If you read the Russian language papers like I do, you’ll notice that the Russian translators don’t verbatim copy what Trump says. Instead, they’ll translate what he’s saying into short, professional, statesman-like statements, instead of what he really says. So instead of “I think nuclear weapons should be way down, and reduced, that’s part of it.”, the Russian audience gets a statement that translates back into English as “I think we’ll start with a substantial decrease in nuclear weapons stockpiles.” When Trump calls something “dumb” or “sad”, the Russian press will have him call it “damned”. “Bad hombres” becomes “armed bandits” and so on.
The Russian support for Trump becomes a lot easier to understand when you realize they’re literally not listening to the same person we are.
It’s really weird to see translation add nuance to a politician’s talk rather than take it away, but this makes so much sense.
But some interpreters, like Tsuruta, are concerned that translating the U.S. president too accurately might reflect badly on them. Alina Cincan, a Romanian interpreter and translator, tells Newsweek: “He’s not known for his eloquence nor his diplomacy. Either you translate exactly what he says which means it’ll be repetitive and sometimes gibberish—you’ll be saying ‘tremendous’, ‘very’ or ‘great’ a lot, or you try to make some sense out of it and ‘beautify’ it. If you choose the former, some may judge your interpreting abilities as poor, whereas if you choose the latter, you’ll make him sound better.”
‘Make America big again’? The headache of translating Trump into foreign languages.
“Most of the time, when he speaks he seems not to know quite where he’s going,” Viennot said. “It’s as if he had thematic clouds in his head that he would pick from with no need of a logical thread to link them.”
She is left with a dilemma: either translate Trump exactly as he speaks — and let French readers struggle with the content — or keep the content, but smooth out the style, “so that it is a little bit more intelligible, leading non-English speakers to believe that Trump is an ordinary politician who speaks properly.”
The thing that I hate about atheism as a movement is that it doesn’t just want to critique the hegemony of Western Christianity, it wants to kill spirituality. There is no joy, there is nothing about it that isn’t founded in a pessimism that sees itself as so self-important that it cannot exist outside of destruction. The face of atheism is a white male disgruntled ex-Christian who decided that if he can’t find joy in religion, then nobody else can. There’s a leftover missionary sensibility to “enlighten” people to atheism that exposes itself as racist, antisemitic, and islamophobic, that’s ultimately not unlike the dominance exerted through colonial Christianity
Atheism ™ is just western Christianity with the spirituality removed.
It is still as proselytizing, self-important and violent as it ever was. It retains the obedience-culture and ‘morality’ of western Christianity dismissing all other thought as primitive and barbaric.
Tom Holland does Rihanna’s “Umbrella” on Lip Sync Battle
I’m literally zendaya reacting like he didn’t have to go that hard and yet..
My interest in Spiderman has gone through the ROOF.
I remember that when P. and I sat down to watch Captain America: Civil War, I said to him – seeing the new Spider-Man actor for the first time – “This kid’s gonna go far.” I had no idea HOW far. 🙂 Dear GOD but he’s terrific.
I regret skipping this video yesterday. This is the best thing I’ve seen all day.
I find myself flummoxed by Guy Ritchie’s bizarre reimagining of the Arthurian legend – giant elephants, kung fu masters, and rodents of unusual size, anyone?
King Arthur: Legend of the Sword US Release Date: May 12th (cinemas)
There are 34 species of tenrec, a small omnivorous animal endemic to Madagascar and parts of mainland Africa. The tenrec female has 29 teats and can give birth to up to 32 young per litter, more than any other mammal. Pictured above: lowland streaked tenrec.
After some research I must report that all of the tenrec furries on DeviantArt are sonic OCs and none of them have an accurate teat count
callanswall 