Reinforcement learning and meta-decision-making

Almost by definition, cognition is flexible: humans (and to an increasing extent, also machines) are not slaves of their environment but act adaptively when novel challenges arise. Critically, adaptive behavior must depend on earlier experience. For example, a tennis player may quickly adapt to playing badminton, padel, or even more remotely related sports such as football. Although the rules, equipment, and playing strategy are not exactly the same, one can profit from the tennis expertise. Thus, a key aspect of cognitive flexibility is to efficiently combine and adapt previously learned knowledge for the current task [1•].

留言 (0)

沒有登入
gif