The role of detours in individual human navigation patterns of complex networks

Abstract

Despite its importance for public transportation, communication within organizations or the general understanding of organized knowledge, our understanding of how human individuals navigate complex networked systems is still limited owing to the lack of datasets recording a sufficient amount of navigation paths of individual humans. Here, we analyse 10587 paths recorded from 259 human subjects when navigating between nodes of a complex word-morph network. We find a clear presence of systematic detours organized around individual hierarchical scaffolds guiding navigation. Our dataset is the first enabling the visualization and analysis of scaffold hierarchies whose presence and role in supporting human navigation is assumed in existing navigational models. By using an informationtheoretic argumentation, we argue that taking short detours following the hierarchical scaffolds is a clear sign of human subjects simplifying the interpretation of the complex networked system by an order of magnitude. We also discuss the role of these scaffolds in the phases of learning to navigate a network from scratch. Everyday life is full of complex networked systems that humans recurringly navigate on a daily basis (e.g., travelling between locations in a city using public transportation). The available navigational datasets 1-4 and models 3-10 considering networked systems mostly target uncovering the average properties of a group of subjects and capture collective human behaviour. Moreover, in terms of human navigation, the existing experiments focus on the dynamic process of learning to navigate, i.e., how people incrementally learn an approximate map of the network. Thus, existing datasets do not have sufficient data or appropriate tracing methods permitting the analysis of long-term individual patterns. Here, we analyse the results of an experiment 11 with human subjects solving navigational tasks in a complex word-morph network. The recorded average of 40.9 timely ordered paths from 259 subjects and more than 200 paths from 9 subjects makes the analysis of individual human navigation patterns possible. In contrast to existing studies, this amount of data enables the inference of characteristics about the steady-state way in which people choose a path between endpoints of a network after they have learned how to navigate the network in their individual way. We argue that this routine navigation process is more valuable to investigate since people use this approach on a daily basis. In the remainder of this study, we refer to this steady-state navigation simply as human navigation. The nodes of the word-morph network, from which we process the navigation paths, are English words that are connected if they differ in only a single letter. In this large and complex network, human subjects are given navigational tasks, i.e., to reach a destination word from a starting word by changing only one letter at a time, while still having meaningful words in intermediate states. Figure 1a shows a sample fragment of the word-morph network an...

A systematic evaluation of assumptions in centrality measures by empirical flow data

“…Sudarshan Iyengar et al (2012) found that the players' solutions are in average 1.7 times longer than the shortest path between the words. When considering only the solutions of experienced players, the solutions' length even decreases to 1.1-1.2 times the optimal solution length (Gulyás et al 2020).…”

“…For the other variant of human navigation in networks where the human themselves travel through the network, there has been a lot of evidence that humans are surprisingly efficient in finding short paths-they are, however, rarely optimal. Sudarshan Iyengar et al (2012) and Gulyás et al (2020) investigated human paths in a word morph game: a player is given two English words, and the player needs to transform the first word into the second word by substituting single letters while intermediate words need to be existing English words. The player's sequence can be seen as a walk through the word network where there is an edge between two existing words if their Hamming distance is exactly one.…”

“…Interestingly, the node AIT is among the 3 most central nodes with respect to all measure variants. Gulyás et al (2020) analyze the navigation paths of this dataset in detail and find that each player builds her own scaffold of the network which she uses in the following for finding short, but not optimal navigation paths. They find that the overlap of the individual scaffolds is very low, hence each player builds her own individual scaffold for effective navigation.…”