Human visual exploration reduces uncertainty about the sensed world

Abstract

In previous papers, we introduced a normative scheme for scene construction and epistemic (visual) searches based upon active inference. This scheme provides a principled account of how people decide where to look, when categorising a visual scene based on its contents. In this paper, we use active inference to explain the visual searches of normal human subjects; enabling us to answer some key questions about visual foraging and salience attribution. First, we asked whether there is any evidence for ‘epistemic foraging’; i.e. exploration that resolves uncertainty about a scene. In brief, we used Bayesian model comparison to compare Markov decision process (MDP) models of scan-paths that did–and did not–contain the epistemic, uncertainty-resolving imperatives for action selection. In the course of this model comparison, we discovered that it was necessary to include non-epistemic (heuristic) policies to explain observed behaviour (e.g., a reading-like strategy that involved scanning from left to right). Despite this use of heuristic policies, model comparison showed that there is substantial evidence for epistemic foraging in the visual exploration of even simple scenes. Second, we compared MDP models that did–and did not–allow for changes in prior expectations over successive blocks of the visual search paradigm. We found that implicit prior beliefs about the speed and accuracy of visual searches changed systematically with experience. Finally, we characterised intersubject variability in terms of subject-specific prior beliefs. Specifically, we used canonical correlation analysis to see if there were any mixtures of prior expectations that could predict between-subject differences in performance; thereby establishing a quantitative link between different behavioural phenotypes and Bayesian belief updating. We demonstrated that better scene categorisation performance is consistently associated with lower reliance on heuristics; i.e., a greater use of a generative model of the scene to direct its exploration.

Compositional clustering in task structure learning

“…Because the CRP is exchangeable, it will have the same 34 predictive error for the sequences ABCDABCD and ADBCDBAC. Order-dependent predictability 35 is beyond of the scope of the current work. 36 We evaluated the CRP on X (n) for n = [1, 100] and compared it to a naïve guess (uniform 37 distribution over M).…”

“…Of course, other decompositions are also possible and useful (and not mutually 599 exclusive). For example, the state-outcome and action-dependent state transition functions of the 600 active inference framework can both be decomposed into "what" and "where" aspects [34,35].…”

“…The state-action value function is used to generate a policy via an epsilon-greedy exploration 34 rule where the action with the highest valueQ(x = x t , a) was chosen with probability 1 ✏ and a 35 random action was chosen with probability ✏ (ties were broken with equal probability). 36 Simulation 1 37 We first simulated the three agents on the same 150 random tasks presented in Simulation 1 of the 38 manuscript.…”

A dual foveal-peripheral visual processing model implements efficient saccade selection

“…One key ingredient to this process is the (internal) representation of 144 counterfactual predictions, that is, the probable consequences of possible hypothesis as they 145 would be realized into actions (here, saccades). Following such an active inference scheme 146 numerical simulations reproduce a sequence of eye movements that fit well with empirical 147 data [MAMF18]. As such, saccades are not the output of a value-based cost function such as 148 a saliency map, but are the consequence of an active strategy by the agent to minimize the 149 uncertainty about his beliefs, knowing his priors on the generative model of the visual world.…”

Active inference, stressors, and psychological trauma: A neuroethological model of (mal)adaptive explore-exploit dynamics in ecological context

“…However, it conflicts with the aim (preference) of avoiding vulnerable self-defensive state outcomes. When the latter aim wins out, heuristically, the stage is set for further epistemic foraging (Mirza et al, 2018).…”