-
-
Notifications
You must be signed in to change notification settings - Fork 990
Add tutorial on Dirichlet Process Mixture Models #880
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Comments
it's easy enough to write a dirichlet-multinomial mixture model (what y'all have been calling dirichlet process, though that terminology is usually reserved for the infinite version), and pretty easy to write the DP in either stick breaking or CRP form. (see, eg, corresponding code in webppl.) however, getting VI to work for mixture models is not entirely straightforward and for infinite models it's still an open problem as far as i know. i have a few ideas, and i think this is a great thing to work on, but having a tutorial in 0.2.1 might be overly optimistic? i may be misinterpreting the scope of what y'all are proposing though! |
Noah, weren't you working with an MIT student on that topic (svi for dpp) ?
Any updates on how this is going?
…On Mar 12, 2018 7:45 AM, "ngoodman" ***@***.***> wrote:
it's easy enough to write a dirichlet-multinomial mixture model (what
y'all have been calling dirichlet process, though that terminology is
usually reserved for the infinite version), and pretty easy to write the DP
in either stick breaking or CRP form. (see, eg, corresponding code in
webppl.) however, getting VI to work for mixture models is not entirely
straightforward and for infinite models it's still an open problem as far
as i know.
i have a few ideas, and i think this is a great thing to work on, but
having a tutorial in 0.2.1 might be overly optimistic? i may be
misinterpreting the scope of what y'all are proposing though!
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#880 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/ABVhL1a2fPIRPqHW-8HQb-7kFHcmpZwoks5tdooQgaJpZM4Sl0xZ>
.
|
hi, any plans on logistic mixture model? |
@zhf459 not from us, but feel free to make one and open a PR! |
See https://forum.pyro.ai/t/variational-inference-for-dirichlet-process-clustering/98 for extended discussion .
It would be nice to make this work with data subsampling.
The text was updated successfully, but these errors were encountered: