Whether the image is mostly noise or mostly data, consistency models go straight to a final result. Image Credits:<\/strong> OpenAI<\/p>\n<\/div>\n

This matters, first, because the pattern in machine learning research is generally that someone establishes a technique, someone else finds a way to make it work better, then others tune it over time while adding computation to produce drastically better results than you started with. That\u2019s more or less how we ended up with both modern diffusion models and ChatGPT. This is a self-limiting process because practically you can only dedicate so much computation to a given task.<\/p>\n

What happens next, though, is a new, more efficient technique that can do what the previous model did, way worse at first but also way more efficiently. Consistency models demonstrate this, though it is still early enough that they can\u2019t be directly compared to diffusion ones.<\/p>\n

But it matters at another level because it indicates how OpenAI, easily the most influential AI research outfit in the world right now, is actively looking past diffusion at the next-generation use cases.<\/p>\n

Yes, if you want to do 1,500 iterations over a minute or two using a cluster of GPUs, you can get stunning results from diffusion models. But what if you want to run an image generator on someone\u2019s phone without draining their battery, or provide ultra-quick results in, say, a live chat interface? Diffusion is simply the wrong tool for the job, and OpenAI\u2019s researchers are actively searching for the right one \u2014 including Ilya Sutskever, a well known name in the field, not to downplay the contributions of the other authors, Yang Song, Prafulla Dhariwal and Mark Chen.<\/p>\n

Whether consistency models are the next big step for OpenAI or just another arrow in its quiver \u2014 the future is almost certainly both multimodal and multi-model \u2014 will depend on how the research plays out. I\u2019ve asked for more details and will update this post if I hear back from the researchers.<\/p>\n<\/p><\/div>\n

<\/br><\/br><\/br><\/p>\n