Papailiopoulos develops MIMO algorithm with GPT-5.6 and Claude Fable 5
Dimitris Papailiopoulos, a University of Wisconsin-Madison professor currently on leave at Microsoft Research, led the work.
Their new algorithm was developed and checked with help from GPT-5.6 and Claude Fable 5, can clean up messy signals in MIMO (multi-antenna) systems, something that used to need super complex math.
Now, it hits top accuracy while running efficiently, which could have implications for wireless communication research, particularly for systems that need to recover information accurately from signals affected by interference and noise.
Papailiopoulos says human effort was key
While it might sound like the AI did all the work, Papailiopoulos explained that human effort was key.
The AI suggested ideas like message passing and bit tweaks, but he spent days refining proofs and fixing errors with GPT-5.6's help.
This shows how AI can be a real teammate for researchers tackling tough problems, especially when things get noisy.