The AI tool, named ‘I-JEPA,’ fills missing sections of photos using background knowledge about the world, which differs from normal generative AI models.
Meta, Facebook and Instagram’s parent company, has now revealed that it will supply researchers with a new ‘human-like’ AI image creation model that can analyse and fill unfinished areas of an image.
The ‘I-JEPA’ AI programme fills in missing elements of photos by using background knowledge about the world. While the I-JEPA may appear to be a standard AI picture generation and fill tool, it is far from it and is touted to be far better.
What distinguishes the I-JEPA is that it incorporates background knowledge about the world when filling in visuals, which improves the output in terms of imagination and does not limit results to the fill. Regular image filling AI techniques, on the other hand, analyse the surrounding pixels of an incomplete image and complete the photo using data from the image itself.
Also Read: Google to face 80% loss in ads business as Europe raising concerns
Meta claims that being inventive and having prior knowledge makes their AI tool far more “human-like.” According to Yann LeCun, a Meta AI scientist, being human-like permits the AI to avoid frequent AI blunders such as producing a hand with extra fingers, and so on.
“For us, it’s way better if the industry standardizes on the basic tools that we’re using and therefore we can benefit from the improvements that others make,” said LeCun.n
Meta is constantly publishing AI models or AI-related models, putting a lot of its bets on artificial intelligence. Meta has included generative AI features into its ad tools, as well as an Instagram tool that would change user photographs based on text commands.
