dfm / dfm/tinygp

Marginalization example - uncertainty

Open
#169 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
349
Forks
35
Avg merge
2d 2h
Merged PRs (30d)
1

Description

In this example

https://tinygp.readthedocs.io/en/stable/tutorials/modeling.html#modeling-numpyro

We marginalize hyperparameters with HMC. In the analysis, we plot quantiles of the posterior mean of y found for each state of hyperparameters in the chain.

I'm a bit confused about what the quantiles of the posterior mean are supposed to represent? In my understanding, they show the uncertainty on the mean of y, rather than the uncertainty on y, since they don't account for the (co)variance.

For example, I could know the mean of y exactly. I still wouldn't know y, though, because of the GP has a covariance as well as a mean.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Read the Modeling with NumPyro section at the linked documentation anchor and inspect how the HMC hyperparameter chain is used to plot quantiles of the posterior mean of y. Determine whether the example needs clarification or a different uncertainty visualization, and document the intended interpretation so the treatment of covariance is unambiguous.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
documentation, machine-learning
Issue type
Documentation
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.