TY - RPRT TI - Online Reward-Punishment Learning from Fixed-Channel Perceptual Event Streams without Environment Rewards AU - Zirong Li PY - 2026 UR - https://arxiv.org/abs/2606.18963 ID - 2606.18963 ER -