Creative foraging and the explore-exploit trade-off in knowledge networks

The scope of knowledge is constantly evolving, due to such factors as environmental changes, cultural evolution, and scientific discovery. Consequently, we are frequently confronted with gaps in our knowledge, compelling us to seek information from available sources. Sometimes the information we seek is easy to find; other times it has yet to be established by others, requiring us to creatively come up with an original perspective. Yet, little is known about how our foraging strategies change depending on the ease with which the information we seek is readily available. We investigated how the need to generate new ideas influences the rate at which individuals explore or exploit existing information. Participants (N=138) answered questions either fully answerable (low-creativity condition) or not fully answerable (high-creativity condition) with information they foraged for on Wikipedia. We created knowledge networks from the foraged information, wherein Wikipedia pages were nodes. The edges linked pairs of Wikipedia pages when they were visited by the participant either sequentially or within the same condition, and were weighted based on the semantic similarity between the pair of pages. This approach allowed us to measure exploration (jumping between disparate pages) and exploitation (viewing closely related pages). In the high-creativity condition, participants were more likely to trade-off between exploration (lower average edge weights) and exploitation (higher average clustering coefficients). This trade-off was associated with responses that were more novel, diverging further from the Wikipedia text, compared to less novel responses. These findings reveal how foraging strategies differ in creative versus non-creative contexts, and provide insight into the processes that underlie learning and scientific discovery.