PFP: A Probabilistic Functional Programming Library for Haskell (2006)
10 days ago
- #functional-programming
- #probabilistic-programming
- #haskell
- Distributions can represent probabilistic events like die rolls or coin flips.
- The `uniform` function creates distributions where each element is equally likely.
- Probabilistic functions (transitions) can be constructed using functions like `choose`.
- Monadic operators can combine probabilistic operations, e.g., rolling a die and then adding one or not.
- The PFP library allows declarative description and evaluation of probabilistic scenarios.
- Example: Calculating the probability of two dice differing by one point yields 27.8%.
- PFP can model problems like the boys/girls riddle, producing results like 2/3 probability.
- Explanation-Oriented Programming focuses on explaining probabilistic reasoning results.
- Papers discuss visual explanations and DSLs for probabilistic reasoning.