The form of uncertainty affects selection for social learning

Area of Science:

• Evolutionary biology
• Behavioral ecology
• Cognitive science

Background:

• Social learning is crucial for adapting to environmental variability.
• Uncertainty, characterized by unknown decisions or outcomes, poses a significant challenge.
• Four key sources of uncertainty are identified: temporal environmental variability, payoff ambiguity, selection-set size, and effective lifespan.

Purpose of the Study:

• To investigate how different forms of uncertainty influence the evolution of social learning.
• To model the interplay between individual and social learning strategies under uncertainty.
• To understand the conditions under which social learning prevails in variable environments.

Main Methods:

• Development of an evolutionary agent-based model.
• Agents learn individually using a softmax decision rule for behavior choice.
• Simulation of multi-armed bandit tasks to represent payoff acquisition under varying uncertainty levels.

Main Results:

• Social learning, specifically vertical and oblique payoff-biased strategies, evolved as a scaffold for individual learning.
• Temporal environmental variability reduced social learning.
• Increased selection-set size promoted social learning, even with frequent environmental changes.

Conclusions:

• Social learning can predominate in highly variable environments by complementing individual learning.
• Effective individual learning aids in overcoming outdated social information, enhancing social learning's robustness.
• The study clarifies the nuanced effects of different uncertainty types on social learning evolution.