Modeling collaborative memory with SAM

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Human Memory

Background:

• Human memory research predominantly uses isolated individuals, neglecting group dynamics.
• Collaborative inhibition is a key phenomenon in collaborative memory, where groups underperform compared to individuals in recall tasks.
• Existing theories lack formal computational modeling guidance.

Purpose of the Study:

• To develop a computational framework for collaborative free recall using the Search of Associative Memory (SAM) model.
• To test if a scaled SAM model can explain collaborative inhibition.
• To provide a unified theoretical basis for both individual and collaborative memory.

Main Methods:

• Scaling the Search of Associative Memory (SAM) model to simulate multiple agents recalling collaboratively.
• Implementing a mechanism where group recall acts as retrieval cues for individual SAM models.
• Analyzing model outputs to identify the emergence of collaborative inhibition.

Main Results:

• The scaled SAM model naturally produced collaborative inhibition.
• The model's results support the "retrieval disruption" hypothesis.
• The framework successfully unifies individual and collaborative memory phenomena within SAM.

Conclusions:

• The Search of Associative Memory (SAM) model provides a unified computational theory for human memory, encompassing both individual and collaborative recall.
• The retrieval disruption hypothesis is computationally supported as an explanation for collaborative inhibition.
• This framework enables future predictions regarding group size, shared knowledge, and false memory dynamics.