Optimizing Outbreak Detection for Real World Networks

Influence Maximization and outbreak detection are two problems that have seen massive research in recent years, for good reason. For example, targeted tweets by well-selected Twitter users can lead to massive reshares, resulting in a quick dissemination of information. Well-placed sensors can allow contamination in a water pipeline network to be detected quickly. Although the two problems may seem entirely different, they are in fact modeled under the same framework. However, it has been known for a while that the problem is intractable in general. Due to this, much research has been done in approximation algorithms in an attempt to find provably good solutions efficiently. We proceed down the same path, and attempt to design efficient outbreak detection algorithms and test it on synthesized network data.