Skip to content

Incorporate AePPL into GaussianRandomWalk #5762

New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

#

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

Jump to bottom
Closed
larryshamalama opened this issue May 14, 2022 · 1 comment
Closed

Incorporate AePPL into GaussianRandomWalk #5762

larryshamalama opened this issue May 14, 2022 · 1 comment
Assignees
@larryshamalama
Labels
enhancements logprob

Comments

@larryshamalama
Copy link
Member

larryshamalama commented May 14, 2022 •
edited by ricardoV94
Loading

A recent PR #5298 added GaussianRandomWalk distribution to v4. It would be possible to use AePPL to obtain the Aesara graph corresponding to desired rv_op and retrieve the correct log-probability graph. The use of AePPL here would be desirable as we will no longer need to implement a corresponding RandomVariable class nor implement the logp method. Moreover, it would generalize the random walk to make use of non-Gaussian distributions.

cc @ricardoV94
@larryshamalama larryshamalama mentioned this issue May 28, 2022
Closed
@ricardoV94 ricardoV94 mentioned this issue Aug 26, 2022
Merged
5 tasks
@ricardoV94
Copy link
Member

Closed via #6072

# for free to join this conversation on GitHub. Already have an account? # to comment
Assignees

@larryshamalama larryshamalama

Labels
enhancements logprob
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

Incorporating AePPL into GaussianRandomWalk larryshamalama/pymc
2 participants
@larryshamalama @ricardoV94