[Sabi Enables AI to Learn Personal Aesthetics from Brain Signals, Generating Faces That Increasingly Match Viewers' Preferences] Beating AI News: Brain-computer interface company Sabi announced a new study called RLNF (Neurofeedback Reinforcement Learning). In the demonstration, an AI face continuously changed, gradually transforming into a form that better matched the viewer's aesthetic preferences. Researchers hope to enable AI to understand each person's preferences from brain signals, without having to repeatedly ask users which image is more attractive. The team first had five participants view a large number of faces while simultaneously recording their brainwaves and eye movements, and then asked them to rate the attractiveness of each face. Using this data, the AI learned how each person's physiological responses would occur when seeing faces they liked or disliked, thus building a model of personal aesthetic preferences. Subsequently, the researchers used this preference data to fine-tune the FLUX.2 Klein 4B image model through reinforcement learning. Each time the AI generates a new batch of faces, the system compares them with previously recorded faces, and provides a reward based on the personal preferences inferred from brain signals. The model continuously adjusts based on rewards, gradually learning to generate faces that better match the aesthetic preferences of specific users, without requiring subjects to score each image individually. Previous research has used EEG feedback to guide AI in generating faces that people like. Sabi, however, uses this feedback for reinforcement learning in image models, allowing personal aesthetic preferences to directly influence the model's generation tendencies. However, the experiment currently only has 5 participants, 2 of whom primarily rely on eye movements to determine their preferences rather than directly identifying them from brain activity. More experiments are needed to prove that EEG can reliably reflect personal aesthetic preferences. [Original Link]