I don’t really see where the ML plays into this. Cool concept though, I’d love to see it ask me if I like such and such and refine the process a bit more to land on something I actually like.
- The colors you see are generated by ML models, not us. If you don't like them, just hit the dislike button and our ML models will soon stop creating them.
- Icon selection based on the description is not an easy problem. You have to understand the semantics of what you're trying to say and match those with an icon's appearance. We're using word embeddings and auto encoders here.
- Initially, a genetic algorithm generates a bunch of stuff but that gets pruned by ML systems. As we get more feedback, that ranking/pruning is going to be more effective.
- Which fonts to combine together is not trivial. Convolutional neural networks come to the rescue here.
- Finally, one of the additional features we are working on is going to generate icons from scratch using GANs.
I hope this answers your question. Let me know if you wanna talk more about how we go about doing everything. Always happy to chat.
May I challenge you to explore these options?
- Not all logos needs colors, actually they should work in black and white as well.
- Your models don't seem to understand meaning or just to be fair, context. A name, tagline is probably too little information to infer a match. For my music tv app, I got 80% notes, keys and other music related icons. Not very imaginative. I usually write down / sketch my first 3 ideas and then throw them away since if it took me 15 minutes it would too for others.
- Why put always an icon there? An icon is a supplement, most successful logos just use a pretty (and expensive) typeface.
- How to do more imaginative typography? I type in The Brand Name and it the engine doesn't even try to change the case of the text.
I would expect to be imaginative, like:
THE..
BRAND
NAME.
Thanks again.