Topic: Bayesian causal inference

Song learning requires birds to interpret sensory evidence, assign credit to changes in behaviour, and reproduce remembered vocal targets. This lecture compares auditory-feedback adaptation, reinforcement learning and imitation as parts of a possible common learning architecture. Incomplete compensation for pitch-shifted feedback is examined through Bayesian causal inference: internal pitch predictions are combined with auditory evidence while feedback unlikely to be self-generated is discounted. Nonlinear pitch trajectories during conditioning motivate an extension of classical policy-gradie

Perceivers benefit from integrating their prior beliefs with sensory signals based on their relative reliability to mitigate the effects of stochastic noise. However, in volatile environments, one should avoid integrating these signals when changes in the signal source render their prior beliefs irrelevant. Bayesian causal inference could offer the brain an optimal strategy for adaptively adjusting its reliance on priors. Mechanistically, it has been suggested that the arousal system modulates the weights between prior beliefs and sensory signals accordingly. We tested these hypotheses using a

Our brains predict future sensory input based on their current beliefs about the world around us, but interpreting prediction errors can be challenging in a volatile environment because they can be caused by stochastic sensory noise or by outdated predictions. Noisy signals should be integrated with prior beliefs to improve precision, but the two should be segregated when environmental changes render prior beliefs irrelevant. Bayesian inference provides a statistically optimal solution to deal with such problems of causal uncertainty [1]. However, the method quickly becomes memory intensive an