Deterministic response strategies in trial-and-error learning
H. Mohr, K. Zwosta, D. Markovic, S. Bitzer, U. Wolfensteller, H. Ruge
This study analyzes human trial-and-error learning strategies, finding that deterministic response models better explain initial learning than Q-learning, with asymptotic performance aligning with Q-learning.
bioRxiv · Feb 1, 2018Animal Behavior & Cognition