A developer's guide to probabilistic programming
summary
Uses F# computation expressions and CSV type providers to turn probabilistic models into first-class code, builds a toy inference engine that does complete enumeration over parameter grids, and fits a mixture-of-Gaussians + Bernoulli model to Stack Overflow 2017 salaries. The dataset has 317 Polish developers with reported salaries; the model returns a most likely mean of $28,400 and a ~25% mistake probability, and the pipeline produces histograms and a working fit.
AI-written summary · about
chapters (4)
- 0:00Start
- 9:00SAMPLING: EXAMPLE MONTY HALL PROBLEM
- 26:12COMPLE ENUMERATION
- 33:32PROBABILISTIC PROGRAMMING IN THE REAL WORLD
related talks
43mHow to look like a statistician (...)
43mThe Power of Probabilistic Programming
29mProbabilistic Programming: What It Is and How It Works
42mBayes and Bool Walk into a Bar
40mProbabilistic Programming for Augmented Intelligence
01h 03mAn Overview of Probabilistic Programming
24mPractical Probabilistic Programming with Figaro
20mProbabilistic Programming with Stochastic Probabilities