Initial proof-of-concept Expectation Propagation

[st--]

This is a very basic version to demonstrate proof-of-concept. See #92 for missing features.

[codecov[bot]]

Codecov Report

Merging #64 (c774892) into master (fb3f410) will increase coverage by 0.74%. The diff coverage is 92.52%.

:exclamation: Current head c774892 differs from pull request most recent head db9b459. Consider uploading reports for the commit db9b459 to get more accurate results

@@            Coverage Diff             @@
##           master      #64      +/-   ##
==========================================
+ Coverage   94.13%   94.88%   +0.74%     
==========================================
  Files           5        6       +1     
  Lines         324      391      +67     
==========================================
+ Hits          305      371      +66     
- Misses         19       20       +1     

Impacted Files	Coverage Δ
src/ApproximateGPs.jl	100.00% <ø> (ø)
src/ep.jl	92.30% <92.30%> (ø)
src/LaplaceApproximationModule.jl	98.07% <100.00%> (ø)
src/TestUtils.jl
src/SparseVariationalApproximationModule.jl	96.25% <0.00%> (+5.77%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update fb3f410...db9b459. Read the comment docs.

[st--]

Why are the tests failing when all I did was merge in master (with #76)? :cry:

[devmotion]

Caused by https://github.com/JuliaDiff/ChainRules.jl/pull/559 it seems (similar to issues in Distributions, DistributionsAD, NNlib etc.). Should be fixed by the new ChainRules release (bugfix was https://github.com/JuliaDiff/ChainRules.jl/pull/562).

[st--]

Caused by JuliaDiff/ChainRules.jl#559 it seems (similar to issues in Distributions, DistributionsAD, NNlib etc.). Should be fixed by the new ChainRules release (bugfix was JuliaDiff/ChainRules.jl#562).

Thanks @devmotion ! it seems to be passing again :)

[st--]

@theogf what do you think of this as a first stab ?