Inferring the structure of sparse propagation networks for the spread of epidemics
Event details
| Date | 01.07.2016 |
| Hour | 10:00 › 12:00 |
| Speaker | William Trouleau |
| Location | |
| Category | Conferences - Seminars |
EDIC Candidacy Exam
Exam President: Prof.Michael Gastpar
Thesis Director: Prof. Matthias Grossglauser
Thesis co-director: Prof. Patrick Thiran
Co-examiner: Prof. Marcel Salathé
Background papers:
Uncovering the structure and temporal dynamics of information propagation.
Reconstructing propagation networks with natural diversity and identifying hidden sources.
The behavior of epidemics under bounded susceptibility.
Abstract:
As the world population becomes more urbanized and interconnected, minor epidemics are more likely to take global dimensions and turn into disastrous pandemics. Modeling the spread of infectious diseases is thus a crucial tool to assist health organizations in controlling the evolution of such epidemics. Recent models tried to improve the state of the art results by taking into account the social network over which the infection propagates. However, observing this network requires an expensive procedure. Therefore, it is usually not accessible to epidemiologists and should be inferred in order to apply advanced epidemic models. The idea of the network inference problem is to reconstruct the network over which an infection spreads while observing only the outcome of a dynamical process propagating over it.
In this proposal, we discuss three recent background works in epidemic modeling and network inference. We first motivate the network inference problem with a study on the effect of network topology on the spread of epidemics. We then investigate two approaches to network inference coming from different academic communities and based on distinct theoretical frameworks. Finally, we show how these studies lay the foundations of our research directions
Exam President: Prof.Michael Gastpar
Thesis Director: Prof. Matthias Grossglauser
Thesis co-director: Prof. Patrick Thiran
Co-examiner: Prof. Marcel Salathé
Background papers:
Uncovering the structure and temporal dynamics of information propagation.
Reconstructing propagation networks with natural diversity and identifying hidden sources.
The behavior of epidemics under bounded susceptibility.
Abstract:
As the world population becomes more urbanized and interconnected, minor epidemics are more likely to take global dimensions and turn into disastrous pandemics. Modeling the spread of infectious diseases is thus a crucial tool to assist health organizations in controlling the evolution of such epidemics. Recent models tried to improve the state of the art results by taking into account the social network over which the infection propagates. However, observing this network requires an expensive procedure. Therefore, it is usually not accessible to epidemiologists and should be inferred in order to apply advanced epidemic models. The idea of the network inference problem is to reconstruct the network over which an infection spreads while observing only the outcome of a dynamical process propagating over it.
In this proposal, we discuss three recent background works in epidemic modeling and network inference. We first motivate the network inference problem with a study on the effect of network topology on the spread of epidemics. We then investigate two approaches to network inference coming from different academic communities and based on distinct theoretical frameworks. Finally, we show how these studies lay the foundations of our research directions
Practical information
- General public
- Free
Contact
- Cecilia Chapuis EDIC