Hand-written MCMC
Write a Metropolis sampler and compare its results with a library such as PyMC.
All pathsDiscover Coursera PlusAfter step 2 After step 3 After step 4
Data & AI · 06Advanced
Probabilistic AI Researcher
Mathematically minded learners drawn to Bayesian methods and graphical models.
Probability, theoretical and applied Bayesian inference, then Bayesian networks and Markov random fields at Stanford.
432hours of coursework
44weeks at 10 h
4steps
14courses
4universities
Target skills
Probability DistributionProbability & StatisticsStatistical InferenceMarkov ModelData AnalysisStatistical AnalysisStatisticsSampling (Statistics)Data ScienceStatistical ModelingBayesian StatisticsStatistical Methods
Skill labels as listed on Coursera.
Target jobs
- Machine learning researcher
- Bayesian data scientist
- R&D engineer
Progression
- 1University of Colorado BoulderIntermediateR, Python, Jupyter Notebooksacademic credit
Stated prerequisites: Calculus through Calculus II (multivariate preferred) and some R or Python experience
Show the 3 courses
Course Level Length Rating (reviews) Probability Foundations for Data Science and AI Intermediate 43 h 4.5 (300) Discrete-Time Markov Chains and Monte Carlo Methods Intermediate 33 h 4.5 (22) Statistical Estimation for Data Science and AI Intermediate 29 h 4.0 (99) - 2University of California, Santa CruzIntermediateR, Microsoft Excel
Show the 5 courses
Course Level Length Rating (reviews) Bayesian Statistics: From Concept to Data Analysis Intermediate 11 h 4.6 (3,233) Bayesian Statistics: Techniques and Models Intermediate 30 h 4.8 (498) Bayesian Statistics: Mixture Models Intermediate 22 h 4.5 (63) Bayesian Statistics: Time Series Analysis Intermediate 19 h 4.3 (18) Bayesian Statistics: Capstone Project Advanced 12 h 4.5 (10) - 3University of PittsburghIntermediatePython, PyMC3
Show the 3 courses
Course Level Length Rating (reviews) Bayesian Inference Fundamentals Intermediate 20 h — Bayesian Regression and Model Selection Intermediate 20 h — Advanced Bayesian Methods and Applications Intermediate 22 h — - 4Stanford University (Stanford Online)AdvancedMATLAB, Octave
Stated prerequisites: Abstract thinking and math skills, programming in at least one language, basic discrete probability (independence, conditional independence, Bayes' rule)
Show the 3 courses
Course Level Length Rating (reviews) Probabilistic Graphical Models 1: Representation Advanced 67 h 4.6 (1,445) Probabilistic Graphical Models 2: Inference Advanced 38 h 4.6 (491) Probabilistic Graphical Models 3: Learning Advanced 66 h 4.6 (305)
With Coursera Plus
This path takes 432 hours, about 10 months at 10 hours a week. A Coursera Plus subscription gives access to most of its courses during that time.
Projects to build
Hierarchical model
Estimate the performance of several schools or hospitals by sharing information across groups.
Diagnostic Bayesian network
Build a small medical Bayesian network and compute the probability of a disease given symptoms.
Free textbooks
Introduction to Probability, 2nd ed.J. K. Blitzstein, J. HwangFree online version
Bayesian Data Analysis, 3rd ed.A. Gelman et al.Free PDF, non-commercial use
Probabilistic Machine Learning: An IntroductionKevin P. MurphyFree PDF, CC BY-NC-ND