Skip to main content Skip to navigation

Calendar

Show all calendar items

DR@W Forum Online: Jianqiao Zhu (Warwick, Department of Psychology)

- Export as iCalendar

Normative models of decision-making optimally transform noisy (sensory) information into categorical decisions, but qualitatively mismatch human behaviour. Leading computational models of behaviour achieve high empirical corroboration by adding task-specific assumptions, deviating from normative principles. In response, we offer a Bayesian approach that implicitly produces a posterior distribution of possible answers (hypotheses) in response to sensory information. But we assume that the brain has no direct access to this posterior, but can only sample hypotheses according to their posterior probabilities. In this sense, we argue that the problem of normative concern in decision-making is integrating stochastic hypotheses, rather than stochastic sensory information, to make categorical decisions. This implies that human response variability arises from posterior sampling rather than sensory noise. Because human hypothesis generation is serially correlated, hypothesis samples will be autocorrelated. Guided by this new problem formulation, we develop a new normatively justified process, the Autocorrelated Bayesian Sampler (ABS), which grounds autocorrelated hypothesis generation in a sophisticated sampling algorithm. The ABS qualitatively explains many empirical effects of probability judgments, estimates, confidence intervals, choice, confidence judgments, response times, and their relationships. Our analysis demonstrates the unifying power of a perspective shift in the exploration of normative models. It also exemplifies the idea that the brain, while fundamentally Bayesian, operates using samples not probabilities, and that variability in human behaviour may primarily reflect computational rather than sensory noise.

Invitation Link

 

Tags: Draw Forum

Show all calendar items