Gall·E 2: AI Artist

Abstract

OpenAI introduced Gall·E 2 lately, enabling their users to get AI-generated pictures based on the meaning of natural languages.

Introducing the AI Artist

On the introduction webpage, OpenAI shows the following picture generated base on the sentence: “An astronaut playing basketball with cats in space as a children’s book illustration.”

I am thrilled to see such AI application as it can be used in many fields. As the pictures can be generated in seconds.
It enables the users to generate high quality pictures with precise elements in the designated pattern and feature, saving users with time, effort as well as painting skills.

But currently, the AI still has its flaws. From the perspective of Natural Language Processing, for example, the sentence explains the picture with singular astronaut and plural cats, but the picture which the AI generated contains the imprecise amount of elements.

Despite of the flaw mentioned above, I think the release of Gall·E 2 is still a very positive signal in the development of AI because it shows the ability of both the understanding, and the creation of arts.

And it’s evolving, according to OpenAI, comparing the pictures generated by its 1^st and the 2^nd generations based on the same sentence: “A painting of a fox sitting in a field at sunrise in the style of Claude Monet.” I would like to applaud to Gall·E 2.

Gall·E 1	Gall·E 2