This repository holds the source code for a research experiment conducted in 2017 called the Supervised Human Neural Network (SHNN).
The paper can be downloaded in pdf format here: Supervised_Human_Neural_Network.pdf
In order to explore how information moves through human networks we use a structured network design from artificial neural networks to impose a deliberate structure on human networks. This acts to quantify how strategic arrangement might affect information transmission between units in a coordinated task. We hypothesize that human networks might perform more accurate classifications simply as a result of how units acting as nodes are structurally con figured.
The experiment was conducted on Amazon's Mechanical Turk platform using a modified version of Otree, a platform for social science experiments.
You can run the experiment for yourself using the following commands:
cd src
pip3 install -U otree-core
otree resetdb
otree runserver
See the Otree documentation for more details.