Google is continuing its efforts to develop Bard. We already know that the conversational agent will be gaining many new features in the weeks to come, and we shouldn’t have to wait long to see an image generator.
As Android app developer Dylan Roussel points out on X.com, Google is planning a Bard update for January 18, 2024. In any case, he has spotted a patch note, not yet officially released, dated this day. The new feature is the ability to create images with Bard.
“It’s easy to use: simply enter a few words to bring your imagination to life,” the document states. The tool allows you to generate multiple visuals from the same prompt, and download the images thus created.
It also states that, initially, this feature is reserved for English prompts. It is not yet known whether there will be any geographical or Google account restrictions.
Bard’s image generator is based on Imagen 2 technology, which converts text commands into images. Google promises results that are “high-quality, photorealistic outputs that are closely aligned and consistent with the user’s prompt.”
One of Google’s goals has been to improve the rendering of human hands and faces, and to generate images free of distracting visual artifacts, two areas where image-generating AIs still have problems.
Google explains that Imagen 2 has been trained on the details of images and associated captions submitted to it. “To help create higher-quality and more accurate images that better align to a user’s prompt, further description was added to image captions in Imagen 2’s training dataset, helping Imagen 2 learn different captioning styles and generalize to better understand a broad range of user prompts,” Google reports.
Thanks to the associations between images and captions, the model is able to better understand the relationship between images and words, “increasing its understanding of context and nuance”.
