Decoding trust: A reinforcement learning perspective