program!

_events/w3.Program.md

@@ -58,6 +58,8 @@ Throughout the event, there will be regular coffee breaks, providing an opportun
 
  - *Algebraic statistics of tree models* (Eliana Duarte):  The goal of this tutorial is to introduce participants to key concepts in algebraic statistics and how these enable us to better understand statistical properties of tree models. One of the main insights in algebraic statistics is that in many cases, statistical models can be defined as the set of solutions to systems of polynomial equations. Understanding algebraic properties of these equations reveal useful properties of the model. The tutorial will start by introducing participants to algebraic concepts such as polynomial rings, ideals and the zero sets that they define via relevant examples in statistics. Next, these concepts will be used to study probability tree models and state the main results about them.
 
+ - *From explanations - to graphs - to geometry - to probabilistic causal reasoning* (Jim Q. Smith): It is not unusual for standard classes of graphical models such as undirected graphical models, Bayesian Networks or staged trees to provide an excellent framework for describing and then analysing their corresponding families of probability models - in particular both their algebraic geometric properties and their causal algebras. But for some applications only a more bespoke family of graphical models can express all the structural information needed to quantify causal inferences. Here I illustrate how it is then still possible to build new bespoke graphs that can then be embellished into other families of statistical to faithfully represent such a domain. Customising models in this way can then lead us to new insightful algebraic geometric descriptions of that domain and in particular more insightful deductions about the causal processes within it, in ways I will illustrate in this talk.
+
 **The event is supported by**