7 March 2023, 18:23

AI can now draw - a new machine that acts on text, audio and visual cues

AI can now draw - a new machine that acts on text, audio and visual cues
Some people think that engineers can do anything. In part they are right. For example, as in the case of the Bot Intelligence Group (BIG) at Carnegie Mellon University. They managed to create a robot very similar to the DALL-E neural network.

This robot got its name in honor of the famous Mexican artist Frida Kahlo - FRIDA. To work, he needs some information about the future project and about an hour of time.

However, this robot is guided not only by text prompts. He can be inspired by an image or a sound. For example, for a recent experience, a photo of Elon Musk and a crying baby were used. FRIDA drew an image of a man with a child's face.
But the researchers didn't stop there. They gave the robot the opportunity to listen to the song. The choice fell Dancing Queen from Abba.

These experiments show the possibilities of AI. But it is unlikely that FRIDA can be perceived as a real artist. This is due to the fact that he cannot be creative. It only acts according to the prescribed algorithm and is guided by prompts. Moreover, its algorithms do not differ from those used in ChatGPT and DALL-E 2.

But in the meantime, FRIDA can create a unique color scheme for each portrait without having to mix the colors yourself.

I wonder what this robot will be able to create in the near future.

To participate in the discussion, please log in.
We use essential cookies for the proper functioning of the website and additional ones to make interaction with the site as convenient as possible. It helps us personalize your user experience as well as obtain analytical information to improve the service.

If you agree to accept all cookies, click "Accept all"; if not, click "Only essential". To learn more, view the Cookie Policy.