That's usually the first question we get from so many folks. From the surface, it doesn't look like an ML problem but here are some of the places where ML is needed. Given a company:
- What color scheme to use? We generate this using machine learning (random forests)
- What fonts combinations to use (We are trying to do this using font embeddings)
- What icon to select given the company name and description (We use word embeddings right now for this. We are also going to use Conv Nets and Auto Encoders to get better icon representations).
- Out of the hundreds of logos we generate using genetic algorithms, which ones to show to user? (This is an image classification problem and we're working on using Conv Nets here).
These are just some of the aspects where we are applying ML. Some of the features are being used and some are being developed.
Please let me know if you have any other questions. Happy to chat more about this.
Faizann20|6 years ago
- What color scheme to use? We generate this using machine learning (random forests)
- What fonts combinations to use (We are trying to do this using font embeddings)
- What icon to select given the company name and description (We use word embeddings right now for this. We are also going to use Conv Nets and Auto Encoders to get better icon representations).
- Out of the hundreds of logos we generate using genetic algorithms, which ones to show to user? (This is an image classification problem and we're working on using Conv Nets here).
These are just some of the aspects where we are applying ML. Some of the features are being used and some are being developed.
Please let me know if you have any other questions. Happy to chat more about this.