Some of the recent image-generating models have this thing where they can fill in the blank parts of images. It’s handy when you want to show them exactly how to give you more of the same. Like these animal emoji. See if you can tell which ones I gave DALL-E2 as examples, and which ones it added.

1st row are standard duck, eagle, owl, and unicorn. After that, it gets weird and hard to identify. There's a pretty good handsome horse, just below a stubby deer with a mushroom for horns, next to which is a teddy bear with a single limpid eye.
Featuring such species as “all comb chicken” and “mushroom deer” and “shopping owl-penguin” and also “cyclops penguin will fight you.”

As opposed to this attempt, in which I used the same text instruction (“A grid of animal emoji”) but didn’t give DALL-E2 any starting example emoji.

Flat cartoon animal heads, most with species unidentifiable, except that they have ears. There are two that might be giraffes, and a pretty hilarious baby elephant, but most are just confusing.
these emoji are suboptimal.

Again, a few examples make a difference. Although it’s interesting that DALL-E2 can make both sets of emoji, but for some reason treats the first kind as more statistically probable when given no other instruction.

The orbs are 3D round beings with two beady eyes and a smiling mouth each. They look both friendly and extremely powerful.
Featuring the three-legged owl-lion, the baby horse-seal, and ORBS

A takeaway here is that internet-trained image generators (and text generators) are trained on so much random stuff that it’s hard to get them to generate the specific kind of thing you’re after. The people making text generators will insert hidden text in the beginning of every conversation to set the stage, and it turns out a similar strategy can pay off for image generation.

The results I’m getting aren’t exactly stellar, even when I seed them with real emoji – image generators are better at vibes than at specific details, and generating small, clearly readable icons is a tough task.

Therefore I can perform a self-roast by asking DALL-E to complete my list of frequently-used emoji.

THe 1st two rows are normal emoji, a thumbs-up, some cakes, a unicorn, some hearts. The last two rows have yelling face emoji, one blue emoji that's all eyes, and some kind of increasing trend graph. Also there is is some kind of raindrop orb with a unicorn horn.

Apparently I have complex thoughts that can only be expressed using an elf with a monobrow.

I also liked “Beverages”.

Top row are a wine glass, squat tumbler, martini glass, and fancy tropical drink. The AI-generated additions include a tropical drink with umbrella, several repeats of the tumbler but with very weird ice, and two martini glasses with intriguing colors of liquids.
Featuring Triangle Ice, Glowing Green Ice, Klein Bottle, and Face Orb With Caterpillar.

And “arrows”

1st row are arrows with messages reading "back", "on", and "soon". The next two have arrows, but the messages say things like Thok, Mlones, Tos, Seusslenes, Talt, and Milty.
Unless you are already familiar with it as food, I recommend not searching for images of “milt”.

These “message emoji” may come in handy?

The 1st row are buttons that say UP!, Cool, New, and Free. The other rows have messages like Mead, Wes, Coo, and Looe, or else icons of hearts, music notes, and swim goggles.
At least some of these would come in handy if your name was Wes and you liked pigeons.

Bonus post: some more examples of extrapolated sets of emoji.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *