stan-dev / stan-dev/stan

Generalize auto-tuning in ADVI to grid search function, for auto-tuning in L-BFGS

Open
#1,780 5 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

algorithm code cleanup
Dominant language
C++
Stars
2.8k
Forks
388
Avg merge
2d 17h
Merged PRs (30d)
15

Description

(There are a number of things I plan to generalize in the code base as I start working on the GMO implementation in Stan. I'm documenting the issues here so I don't forget.)

See title. We can use the same grid search to optimize hyperparameters for many things in Stan. The easiest extension at the moment is to make the grid search a stand-alone function, for auto-tuning the initial step-size parameter in L-BFGS.

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

Start by locating the existing ADVI auto-tuning grid-search implementation and the L-BFGS initial step-size code. Identify the shared behavior that can become a standalone function, then verify that it supports auto-tuning in both contexts without changing their intended behavior.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
machine-learning
Issue type
Refactor
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.