How social learning enhances-or undermines-efficiency and flexibility in collective decision-making under uncertainty

Area of Science:

• Cognitive Neuroscience
• Computational Social Science
• Collective Intelligence

Background:

• Balancing efficiency and flexibility is crucial for collective decision-making amid rapid societal changes.
• Two computational algorithms for social learning, value shaping (VS) and decision biasing (DB), have been proposed.
• VS uses others' choices as rewards, while DB relies on personal experience for valuation updates.

Purpose of the Study:

• To explore the interactive dynamics and group-level consequences of VS and DB.
• To examine collective performance in uncertain and dynamically changing environments using computational models.
• To understand how social learning strategies impact collective intelligence.

Main Methods:

• Developed computational models of VS and DB within a reinforcement learning framework.
• Conducted agent-based simulations to analyze collective decision-making.
• Performed evolutionary analyses to assess the stability and coexistence of learning types.

Main Results:

• A trade-off exists: VS provides efficiency in stable contexts, while DB offers adaptability in volatile environments.
• These differences are magnified in larger groups and under strong majority influence.
• Both VS and DB can coexist stably, enhancing overall group performance.

Conclusions:

• Social learning strategies significantly influence collective intelligence, with potential to both enhance and impair group performance.
• The findings suggest design principles for resilient collective decision systems in human and AI societies.
• Understanding these social learning dynamics is key for navigating complex, changing environments.