What does it imply for an AI to be artistic?
Final yr, I wrote an article about measuring creativity in Giant Language Fashions (LLMs) utilizing a number of word-based creativity checks.
Since then, AI has developed quickly and is able to processing and creating each textual content and picture. These fashions, typically known as “Multimodal Giant Language Fashions” (MLLMs), are extraordinarily highly effective and have superior talents to know complicated textual and visible inputs.
On this article, I discover one solution to measure creativity in two well-liked MLLMs: OpenAI’s GPT-4 Vision and Google’s Gemini Pro Vision. I take advantage of rebus puzzles, that are phrase puzzles that require combining each visible and language cues to unravel.
Creativity is extraordinarily multi-faceted and troublesome to outline as a single trait. Due to this fact, on this article, I intention to not measure creativity typically, however to guage one very particular facet of creativity.
Be aware [modified from my earlier article]: These experiments intention to not measure how artistic AI fashions are, however moderately to measure the extent of artistic course of current of their mannequin generations. I’m not claiming that AI fashions possess artistic considering in the identical method people do. Slightly, I intention to indicate how the fashions reply to explicit measures of artistic processes.
A rebus puzzle is an image illustration of widespread phrases or phrases. They typically contain a mixture of visible and spatial cues. For instance, under are six examples of rebus puzzles (solutions are on the finish of the article).