Applicative bisimulation for higher-order probabilistic languages
We will recall the foundations of Abramsky's applicative bisimulation for the pure lambda-calculus, and present its generalisations to probabilistic lambda-calculi (for both discrete and continuous probabilities). We will conclude by some perspectives on how to make applicative bisimulation quantitative, leading to ideas of distances between programs.