List of projects for Fall 2018 - UBC-Stat-ML/blangSDK GitHub Wiki

Phase 1: learning and using Blang to do data science

I recommend to do at least a bit of those before moving to phase 2.

• Tutorials around specific models or concepts such as plates, etc
• Integrating implemented models into the Benchmark bonanza repo
• Documenting common errors and how to fix them (FAQ in wiki)

Phase 2: becoming a core contributor

Don't be shy to create pull requests into the blangSDK repo. I will provide feedback and we can iterate.

• Performance improvements to get on par with state-of-the-art. E.g. improvements on the simplex sampler (Issue #101). Profiling to figure out where time is spent.
• Add standard distributions to the SDK with associated tests; see Wikipedia's list
• Tools to streamline data-analysis: automate summary reporting (ESS, CIs, etc), intelligent plotting
• Implement an approximate Jeffrey Prior distribution

Phase 3: more advanced, long term projects

• Get involved in the cancer genetic module currently under development, e.g. by adding pyclone to it
• Create and lead the development of a phylogenetics module
• Create and lead the development of a Bayesian non-parametric module
• Create and lead the development of a spatial module