Internal BACS Workshop on Bayesian Models and Contraction Theory
This workshop is organized in two parts:
- One part on Contraction Theory, aimed at discussing the state of the art in nonlinear contraction theory as it pertains to neuroscience applications, and at exploring in some detail the recent tools of stochastic contraction in relation to Bayesian ideas and questions studied in BACS. The morning will be devoted to presentations of existing results, while the afternoon should leave room for extended discussions.
- One part on Bayesian Models, which aims at making an inventory and discussing the theoretical Bayesian Models used in the BACS community. Participants will be asked to present their models and possibly identify a couple of difficult points to be the basis of discussion. We hope to have room for extended discussions, with at least one hour for each model presented. Each participant will have to describe the context of their problem, the variables appearing in the model, the Bayesian Network or Decomposition for the dependency structure, the probability distributions used and their learning mechanisms, if any, and the inference techniques and possible difficulties. Each interested BACS member is asked to provide a one or two-page description of their model according to the guidelines above, and/or scientific papers including such a description, if available.