Implement ordinal stopping-ratio distribution #963
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Formula from the Handbook of Differential Entropy, J.V. Michalowicsz, J.M. Nichols, F. Bucholtz p. 113, 114
Co-Authored-By: Srinivas Vasudevan <[email protected]>
Co-authored-by: Srinivas Vasudevan <[email protected]>
Co-authored-by: Srinivas Vasudevan <[email protected]>
Co-authored-by: Srinivas Vasudevan <[email protected]>
…n mean and variance
Replace scale and concentration by the underlying logistic parameters : loc and scale. !!!WARNING: some numpy tests fail for an unknown reason and still need to be fixed!!!
Co-authored-by: Srinivas Vasudevan <[email protected]>
Co-authored-by: Srinivas Vasudevan <[email protected]>
Optimize import and fix names in _params_event_ndims
… at present (computed Boolean masks). Add jax_omit_deps to allow this. PiperOrigin-RevId: 318536818
END_PUBLIC *** Original change description *** BEGIN_PUBLIC Disable hidden_markov_model_test on JAX. PiperOrigin-RevId: 318537913
For rewrite.py, this meant just adding more runtime checks to convince PyType that we're doing everything properly. The final code certainly has better errors. For fun_mcmc_lib.py, there's nothing we can do as nests cannot be described precisely by the type system, so we switch to regular casts. PiperOrigin-RevId: 318546801
…interaction not yet well supported by tfp.jax. PiperOrigin-RevId: 318885402
PiperOrigin-RevId: 318885822
Also revert horseshoe_test now that JAX uses a more numerically stable log1p. PiperOrigin-RevId: 318899551
PiperOrigin-RevId: 318904656
….TransformedDistribution` in tfp.sts. PiperOrigin-RevId: 318922723
PiperOrigin-RevId: 318924401
|
I think I messed up to rebase the latest commits from master into my branch. I guess it makes more sense to open another PR and close this once we finished the reviews? |
|
Sorry, forgot to ping @srvasude . |
|
This will be impossible to review until we sort out the base (it's showing changes in 361 files). Feel free to send another PR and close this if it's easier for you. |
|
Yeah, you're right, sorry. I'll close this and open up another one. |
copybara-service bot
pushed a commit
that referenced
this pull request
Dec 8, 2020
Hello, this PR again implements the stopping ratio logistic distribution which was already discussed and reviewed [here](#963). @srvasude , sorry for closing the other PR. I worked in your requested changes other than the transpose that needs to be done when sampling. Thanks. Cheers, Simon COPYBARA_INTEGRATE_REVIEW=#990 from dirmeier:stopping_ratio 0743b54 PiperOrigin-RevId: 346376187
jburnim
pushed a commit
to jburnim/probability
that referenced
this pull request
Dec 22, 2020
Hello, this PR again implements the stopping ratio logistic distribution which was already discussed and reviewed [here](tensorflow#963). @srvasude , sorry for closing the other PR. I worked in your requested changes other than the transpose that needs to be done when sampling. Thanks. Cheers, Simon COPYBARA_INTEGRATE_REVIEW=tensorflow#990 from dirmeier:stopping_ratio 0743b54 PiperOrigin-RevId: 346376187
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hello,
this PR implements the sequential stopping-ratio distribution, an ordinal distribution where categories can only be reached one after another. Solves this issue.
Cheers,
Simon