2024/07/18: 9h30
Speaker : Pierre-Henri Wuillemin (LIP6, Sorbonne Université)
Title : Probabilistic Graph Models and Causal models for AI ( Modèles probabilistes graphiques, modèles causaux )
Abstract: La modélisation probabiliste est souvent considéré comme complexe, nécessitant un nombre exponentiel de paramètres, même dans le cas discret. Dans cette présentation, nous nous attacherons à décrire et à motiver les modèles graphiques probabilistes. Après avoir abordé quelques propriétés importantes, nous montrerons en quoi ces mêmes modèles graphiques peuvent servir de support à une analyse quantitative causale, ramenant dans le champs des mathématiques et de la statistique une notion qui a longtemps été considérée pour le moins peu formalisable. (Cette présentation se basera principalement sur les travaux de Judea Pearl, prix Turing 2011)
Pause Cafe : 10:30 — 11:00
11h00
Speaker : Christophe Gonzales (COALA, LIS)
Title : A brief overview of decision under uncertainty
Abstract: Decision under uncertainty is pervasive in artificial intelligence. The goal of this tutorial is to review some popular decision models and highlight their connections as well as the situations in which they can be applied. We will start with the expected utility model (EU) and show the properties it relies on. Then, we will present decision trees, that represent sequential decision problems, and show that the aforementioned properties allow for an efficient algorithm to solve them. Interpreting these trees differently will lead to another more efficient model called an influence diagram. The EU model is known to have severe limitations and, in some situations, more general decision models are needed. Based on a rephrasing of EU, we will present the more general rank dependent utility model (RDU). We will also show the issues of RDU w.r.t. sequential decision making. Another path to generalize EU is to substitute probabilities by other models, notably belief functions. We will show that the EU’s properties presented at the beginning of the talk can also be applied to belief functions, hence resulting in the belief expected utility (BEU) model. We will conclude this talk, mentioning briefly other popular decision models.