adaptive

adaptive copied to clipboard

AverageLearners in a BalancingLearner don't work well

2

In a `BalancingLearner` where one of the learners returns `contant+-eps` (where eps < 1e-15) the points are not correctly balanced over the learners. The one that returns 0 (or a...

basnijholt

(Runner) add live-info to runner in a terminal

2

## ([original issue on GitLab](https://gitlab.kwant-project.org/qt/adaptive/issues/70)) _opened by Jorn Hoofwijk ([@Jorn](https://gitlab.kwant-project.org/Jorn)) at 2018-07-05T17:59:49.217Z_ I often run adaptive in a terminal, and to me it seems very useful to print a status...

basnijholt

I am using `dask` where the number of cores is adaptively scaled, `overhead` now reports a negative number, which is definitely wrong. 

basnijholt

Literature to look at

4

This issue is like a "to-do list" and these are suggestions people made. * [Ruppert's algorithm](https://en.wikipedia.org/wiki/Ruppert%27s_algorithm) as suggested [here](https://www.reddit.com/r/dataisbeautiful/comments/b8vv2p/i_wrote_a_python_package_to_do_adaptive_sampling/ek0qrfl/?context=3) * [Wavelet compression](https://www.youtube.com/watch?v=txk-VO1hzBY&feature=youtu.be) as suggested [here](https://www.reddit.com/r/dataisbeautiful/comments/b8vv2p/i_wrote_a_python_package_to_do_adaptive_sampling/ek0rl02/?context=3) * compare performance with hyperopt...

basnijholt

document loss function to find the band gap

4

Many people have asked me how to do this. I've recommended them to use the following loss function: ```python @adaptive.learner.learner1D.uses_nth_neighbors(0) def abs_min_log_loss(xs, ys): from adaptive.learner.learner1D import default_loss ys = [np.log(np.abs(y).min())...

basnijholt

documentation issues

6

- [ ] [documentation of stable is broken](https://adaptive.readthedocs.io/en/stable/tutorial/tutorial.BalancingLearner.html) - [ ] the page loads very slowly, see https://github.com/rtfd/readthedocs.org/issues/5515 - [ ] we need to be able to display the documentation...

basnijholt

Specify an API for defining the scale of point

1

## ([original issue on GitLab](https://gitlab.kwant-project.org/qt/adaptive/issues/123)) _opened by Jorn Hoofwijk ([@Jorn](https://gitlab.kwant-project.org/Jorn)) at 2018-11-19T10:33:48.135Z_ Instead of having the loss function deal with scaling, we may pass the scaled points into the loss...

basnijholt

Consider allowing current learners to work with functions defined on integers

4

## ([original issue on GitLab](https://gitlab.kwant-project.org/qt/adaptive/issues/67)) _opened by Anton Akhmerov ([@anton-akhmerov](https://gitlab.kwant-project.org/anton-akhmerov)) at 2018-07-02T14:11:30.824Z_ I'm not actually sure how to do it, but for instance I could imagine studying how things depend...

basnijholt

Add averaging 1D learner

4

## ([original issue on GitLab](https://gitlab.kwant-project.org/qt/adaptive/issues/6)) _opened by Joseph Weston ([@jbweston](https://gitlab.kwant-project.org/jbweston)) at 2017-07-25T15:52:41.859Z_ Add a learner that can learn a function that returns a random variable. This is like the 1D...

jbweston

## ([original issue on GitLab](https://gitlab.kwant-project.org/qt/adaptive/issues/130)) _opened by Bas Nijholt ([@basnijholt](https://gitlab.kwant-project.org/basnijholt)) at 2018-12-07T19:57:55.908Z_ I closed a bunch of issues that I think could be closed, however, there are some more that...

basnijholt