This ain't your daddy's probability monad!

We'll meet at 19:00 at the pizzeria Unikat.

The talk starts between 20:00 and 20:30 in the Math building in room M101.

Lars will talk about probability monads and their extensions with the treatment of time and parallelism.

Probability monads are a very nice example of monads. They make it easy to simulate situations where randomness is involved, like dice and card games or the infamous "Monty Hall" problem. They serve as a nice illustration of how monads can be implemented and how one can work with them - using type classes, direct implementations or free monads with different interpretations.

For example, we can sample from a probabilistic computation, but we can also get exact probability distributions by using different interpretations and implementations of the same abstract interface.

Things become more interesting, but also trickier, once in addition to probabilities, the concept of time (or delay) is added, especially in the context of parallel processes.

The talk will start simple by revisiting standard probability monads and then progress to process monads that can model information flow in complex networks, where signals have random delays and can even get lost.