可引用的發言段落Citable transcript turn
#24705
講者：Speaker:
They could create a translation on memory or lexicon, and then yet another artificial intelligence can take that and try to create some 3D models, pictures and try to find relevant images that corresponds to what we worked on.