Abstract
The sixth seminar looked at probabilistic languages. These are simple functional languages such as PCF to which randomness is added, for example a primitive that evaluates to 0 with probability ½ and 1 with probability ½. These languages are used in particular to study Bayesian inference. The speaker developed denotational semantics for the probabilistic PCF language using probabilistic coherent spaces. She drew links with linear logic and in particular its semi-polarized fragment, which corresponds, in Curry-Howard's sense, to Levy's call-by-push-value calculus.
