The L1Distance is often numerically unstable
Open
@dfm is already working on this.
Since Oct 28, 2022.
enhancement
- Dominant language
- Python
- Stars
- 349
- Forks
- 35
- Avg merge
- 2d 2h
- Merged PRs (30d)
- 1
Description
Is this a bug? Should we change the default distance for the Matern kernels?
For example, the following:
import numpy as np
from tinygp import GaussianProcess, kernels
random = np.random.default_rng(48392)
X = random.uniform(-1, 1, (100, 2))
kernel = kernels.Matern32()
print(GaussianProcess(kernel, X, diag=0.01).solver.normalization())
returns NaN, but replacing the kernel with:
kernel = kernels.Matern32(distance=kernels.L2Distance())
works…
See also #115
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.