Detecting epidemic sources in presence of misinformation


When an epidemic spread in a network, being able to detect its source, i.e., the node that initiated the propagation, is extremely useful. For example, this could help to develop appropriate containment measures or to prevent future epidemics.

In this project, we look at an innovative and lifelike scenario in which we do not a priori have any information about the state of the network nodes. Instead, we investigate the source identity by directly going on the terrain, i.e., by sequentially querying some nodes. For example, we could ask them who they were infected from or when they became infected. However, in a realistic scenario, the nodes do not always provide correct information: They could be uncertain about what exactly happened or they could be malicious, i.e., they could purposely give some wrong information. How can we then use the information available to detect the source in the shortest time and with the highest accuracy?