What is the model in model-based reinforcement learning?
Human intelligence has long been explored in the context of complex, sequential decision making tasks like chess and go. While there has been much theoretical and empirical support for model-based reinforcement learning (MBRL) accounts of human behavior in these kinds of tasks, much of the existing work in cognitive science has been limited to MBRL algorithms with relatively simplistic representations of task dynamics. In this paper we argue that these simple representations fail to capture a...
Casus belli: a coalitional theory of moral outrages
The killing of Trayvon Martin. The Orlando massacre. The attack on Pearl Harbor. These grave events are dramatically different from one another on surface properties but are brought together by a feature of our psychology tailored to solving the problem of collective deterrence. An extension of the Recalibrational Theory of Anger, the coalitional theory of moral outrages suggests that a suite of motivational goals is brought to bear on the unique problem of collective deterrence. Unlike the deterrence of one...
Inferring "Us" and “Them” through Latent Structure Learning
Humans form social coalitions in every society on earth, yet little is known about how we learn and represent social group boundaries. Previous research has focused on group inference through explicit information (e.g., team memberships and visual cues). How do we accumulate information from the environment to infer social group membership when such labels are not explicitly available? We derive predictions from a computational model of latent structure learning to move beyond explicit category labels...
Evidence that Parallel Distributed Networks Dissociate Episodic and Social Functions Within the Individual
Association cortex is organized into large-scale, distributed networks. One such network, often called the default network, has been linked to diverse forms of internal mentation, opening debate about whether shared anatomy supports multiple forms of cognition or subtle distinctions in cortical organization have yet to be resolved. Using recently developed procedures for high-resolution analysis in ten individuals, we probed whether multiple tasks from...
What if everybody did that? Universalization as a mechanism of moral judgment
Some people feel that it is morally wrong to not vote, but why? From a utilitarian perspective, one vote fewer often makes no practical difference; from a deontological perspective, there is often no rule or law that mandates voting. Current theories of moral psychology struggle, then, to explain why anybody would consider it morally wrong. A very natural answer to this problem comes to mind, however: We ask ourselves, “What if everybody did that”? My aim in this talk is to formalize...
Model-free and model-based learning processes in the updating of explicit and implicit evaluations
Evaluating stimuli along a positive–negative dimension is a fundamental computation performed by the human mind. In recent decades, research has documented both dissociations and associations between explicit (self-reported) and implicit (indirectly measured) forms of evaluations. Together, these two forms of evaluation are central to organizing social cognition and drive behavior in intergroup relations, consumer choice, psychopathology, and close...
Why are humans the only ape with a prominent external nose? To answer this question, we must ask: why do animals have noses? A major selective force in the evolution of olfaction is its role in spatial navigation; navigation and olfaction are intimately connected at the neural level across the animal kingdom, even in humans. In vertebrates, this is reflected in the robust and conserved circuitry linking the hippocampus, a navigational center, with...