Modern computer vision algorithms employ “zero-shot learning,” where a model can recognize objects without prior training on them. By leveraging auxiliary information like text...