Final Report Summary - TRANSFER-LEARNING (Transfer Learning within and between brains)
Decision theory based on ultra-rational decision makers has failed to provide a satisfactory understanding of highly complex, and interdependent human behavior. Our project aims at investigating the neural bases of human adaptive behavior in private and social environments. We use both computational and neuroscientific methodologies to provide accurate models of learning in private and social settings. The main goal is to develop a model of adaptive learning based on three basic principles: (1) The observation of the outcome of the un-chosen options improves the learning process. Thus, individuals learn by comparing the outcome of chosen and un-chosen actions (i.e. regret learning). (2) Learning can be transferred from one situation to another. Transfer learning (or generalization) refers to the improvement of performance in one task due to the knowledge acquired in a related task. This is a crucial aspect of learning and it answers the question about how acquired knowledge and learned rules in one task can be used in novel tasks. (3) Learning can be transferred from one agent to another (i.e. social learning). Learning by observation and other forms of transmission of knowledge and rules between individuals could improve learning and generalization. The results of the of our project show: (i) the adaptive function of emotions such as regret in learning and decision making; (ii) evidence of transfer learning; and (iii) highlight the fundamental role of observational learning in social settings. The results of our research could contribute to building a neural, behavioral, and theoretical foundation for transfer learning in private and social settings. Such foundations will eventually result in a better understanding of how humans respond to environmental changes, and with a better understanding at our disposal, society should be better able to cope with the inherent instability of the modern economy. Beyond their impact on future economic policies, the results of the current project should be generally exploitable by all those interested in and concerned by the way in which individuals and groups learn to adapt in a variety of circumstances.