LNC2, ENS
Constraint and advantage: the two faces of noise in decision making under uncertainty
Human decisions made under uncertainty are not only subject to biases, but also to internal noise. I will argue that this noise has a dual impact on decision making. First, noise acts as a constraint that shapes how humans solve decision problems. Second, in artificial neural networks, noise shapes the selection of abstract and resilient task representations that resemble those measured in human brain activity.