Rohan Paul reviews RLNF: using brain signals as rewards
Rohan Paul introduces RL with Neural Feedback (RLNF): while viewers look at generated images, arousal signals are read directly from the brain, and a simple embedding plus k-NN scoring is used as the reinforcement learning reward. He considers this direction very natural—human preferences already exist in the brain; RLHF turns them into clicks, while RLNF connects brain signals directly to the reward.
View original postAll that said, I'm generally excited about this new paradigm. It gives us a whole new axis to scale AI models - towards personalized responses. Which is more impactful to large scale adoption than just intelligence. If AI can understand you, it can make things more beautiful for you. We are hoping to help make a more beautiful world for everyone.
View original
All that said, I'm generally excited about this new paradigm. It gives us a whole new axis to scale AI models - towards personalized responses. Which is more impactful to large scale adoption than just intelligence. If AI can understand you, it can make things more beautiful for you. We are hoping to help make a more beautiful world for everyone.
Body
this is just so next level for personalized AI.
arousal signals scored by a simple embedding plus k-NN reward. RL with Neural Feedback (RLNF), reads a viewer’s arousal signals directly from the brain, while they look at a generated image and uses them as the RL reward.
this direction looks so natural now. Human preference anyway lives in the brain, RLHF routed it through a click, and RLNF routes it from brain signals straight into the reward.