Surfing Uncertainty: Prediction, Action, and the Embodied Mind

In the citing paper

“In addition, although Broca's area typically activates during performance of IS tasks [43], there are reasons for caution in assuming that neural signatures elicited by task instructions will align with IS that arises more spontaneously [17].”

In the citing paper

“Hierarchical predictive processing (HPP) offers powerfully explanatory implementational and algorithmic details for active inference (Clark, 2016), providing a single mechanism for both perception and action.”

In the citing paper

“According to these theories, predictions and incoming sensory input each have an associated precision (inverse variance) reflecting their confidence or reliability.”

In the citing paper

“a mechanism known as ‘precision weighting’ modulates the impact within the overall system that sets of errors have: high weighting drives learning and further engagement, while low weighting renders error units relatively impotent”

In the citing paper

“The AIF aims to deliver a process theory explaining how agents remain within expected bounds, minimizing uncertainty within volatile environments”

In the citing paper

“In AIF terms, the brain encodes a probabilistic generative model of the causes of its sensations, which it leverages to generate action”

In the citing paper

“If predictions fail to match the incoming sensory data, ‘prediction errors’ are generated that propagate upward through the hierarchy of the generative model, driving action and revised prediction.”

In the citing paper

“Rather than passively waiting for incoming sensory information, systems are proactive, generating ‘top–down’ predictions about future states”

In the citing paper

“In predictable situations, organisms may rely on automatized reflexive responses (in which case, the biologically viable predictions are made through natural selection and embodied in the phenotype; see Clark, 2016).”