The quest for simplicity in human learning: Identifying the constraints on attention

Area of Science:

• Cognitive Science
• Computational Neuroscience
• Psychology

Background:

• Human cognition operates under severe constraints, with limited sensory input reaching awareness and even less being stored in memory.
• Selective attention is often viewed as an optimization process, directing cognitive resources to information that minimizes errors.
• Understanding these constraints is crucial for modeling learning and decision-making.

Purpose of the Study:

• To investigate the impact of attention constraints on category learning.
• To develop and test computational models of attention with varying levels of constraint.
• To identify underlying computational principles governing attention during learning.

Main Methods:

• Developed computational models incorporating different constraints on attention during learning events.
• Fitted these models to behavioral (choice) and eye-fixation data from established category-learning datasets.
• Analyzed data from a dynamic categorization task to assess model performance in a more complex scenario.

Main Results:

• Found consistent evidence for a "simplicity" computation, where attention is minimized to the fewest necessary information dimensions.
• Observed a
• competition
• mechanism, where different information dimensions vie for attentional resources through lateral inhibition.
• Models incorporating these constraints successfully explained observed learning patterns in the data.

Conclusions:

• Attention constraints significantly shape the learning process.
• The principles of simplicity and competition appear to be fundamental computational mechanisms governing selective attention in learning.
• These findings offer a refined understanding of how the brain optimizes information processing under resource limitations.