,
5 tweets,
2 min read
Read on Twitter
Just up on arXiv from @ChrisWLynn "Human information processing in complex networks" arxiv.org/abs/1906.00926 w/@ari_e_kahn & L Papadopoulos, combining information theory & cognitive science to quantify the information that humans receive from complex communication systems. 
Traditional methods in information theory treat the production of information as an inherent property of a system. In contrast, here we consider the human perspective by accounting for the subtle, yet critical, fact that information depends on the expectations of a receiver.
By incorporating human expectations, we develop an analytic framework to study the information perceived by a human observer, which we motivate with new human experiments (our first main result).
Applying our framework to several real-world communication networks, we find, remarkably, that all of them share two information properties: (i) they communicate large amounts of information & (ii) they do so efficiently by remaining close to human expectations (2nd main result).
Additionally, we show that this rapid and efficient communication emerges in networks that are simultaneously heterogeneous and modular, the two defining features of hierarchical organization (our third main result).
