Coder Social home page Coder Social logo

mingzhang-yin / sivi Goto Github PK

View Code? Open in Web Editor NEW
51.0 5.0 14.0 756 KB

Using neural networks to build an expressive hierarchical distribution; A variational inference method to accurately estimate the posterior uncertainty; A fast and general method for approximate Bayesian inference. (ICML 2018)

Home Page: http://proceedings.mlr.press/v80/yin18b/yin18b.pdf

License: MIT License

Python 47.62% MATLAB 50.73% C 1.65%
variational-inference approximate-bayesian-inference variational-autoencoder uncertainty-quantification implicit-flow implicit-modeling mcmc

sivi's People

Contributors

mingzhang-yin avatar

Stargazers

avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar avatar

Watchers

avatar avatar avatar avatar avatar

Forkers

mingyuanzhou mylesbartlett yh-utmsb siamakz satyam-cyc yuguangyue jingzhang1 cognoscentai wangyongguang namini94 zhougroup alfieriek pkulwj1994

sivi's Issues

Normalizing constant for q(z|ψ)

Hello

Thank you for posting the source code for your experiments! While viewing it I noticed that computation of the density q(z|ψ) at SIVAE.py#L113 seemingly lacks the normalizing constant that includes the changing variance term.

Am I missing something? Is it cancelled out somehow?

What is the role of `WU`

Hello, I have two doubts reading your code. 1. what is the variable WU in the VAE code? I noticed that it is assigned by warmup and will increase from 0 to 1 as the epoch increases. Means that the weight of the likelihood gradually decreases with training? Is there any consideration behind this?

SIVI/SIVAE.py

Line 126 in c457d09

loss_iw0 = -tf.reduce_logsumexp(log_lik_iw+(log_prior_iw-log_H_iw)*WU,1)+tf.log(tf.cast(K,tf.float32))

2.One more question about this loss, I noticed that there is a batch count multiplier missing. Is that right？
image

Thanks!

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.