Random walks on semantic networks can resemble optimal foraging

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Semantic Network Analysis

Background:

• Memory retrieval often yields clusters of semantically related items.
• Previous work proposed this reflects optimal foraging strategies in information retrieval.
• This optimal foraging model suggests strategic switching from depleted information clusters.

Purpose of the Study:

• To investigate alternative explanations for clustered memory retrieval.
• To explore whether random walk processes on semantic networks can replicate observed retrieval patterns.
• To challenge the necessity of strategic search in explaining memory recall phenomena.

Main Methods:

• Constructed a semantic network from human word-association data.
• Simulated random walk processes on this semantic network.
• Analyzed the emergent retrieval patterns from the random walks.
• Compared simulation results with empirical findings on clustered memory recall.

Main Results:

• Random walks on the semantic network produced retrieval patterns similar to optimal foraging.
• Clustered retrieval emerged when semantically related items were spatially proximate in the network.
• The findings suggest undirected search processes can account for the observed phenomena.

Conclusions:

• Memory search may not require strategic decisions akin to optimal foraging.
• Random walk models offer a viable alternative explanation for clustered memory retrieval.
• Different cognitive processes can yield similar behavioral outcomes depending on the underlying representational structure.