144 questions
0
votes
0
answers
16
views
Noiseless Gausian Proces regression not crossing the training points
I'm trying to perform a Gaussian Process Regression on the following series of data points (X and Y):
X Y
0.00001 -6.8200000000000003
0.00014 -6.8073846153846160
0.00042 -6.6818333333333326
...
- 1
0
votes
0
answers
47
views
How do I create a multitask GPyTorch model with a user-specified noise covariance matrix?
I've implemented standard homoskedastic multitask Gaussian process regression using GPyTorch as follows:
class MyModel(gpytorch.models.ExactGP):
def __init__(self, X, Y, likelihood):
super(...
0
votes
2
answers
97
views
Specify minimum surface constraint for sklearn GPR prediction
I'm trying to use sklearn.GaussianProccessRegression to fit data, then estimate new points using the regression. The predictions need to be positive (preferably above a specified minimum).
I'd like to ...
- 91
2
votes
0
answers
56
views
sklearn Gassian process with multiple targets vs gpytorch
I am performing GaussianProcess regression of a function-to-function relationship. I have N different functions sampled at discrete points p to build an input X whose shape is N by p and these map row ...
- 837
0
votes
0
answers
17
views
Gaussian Process that models the uncertainty of the samples, not of the posterior
I have some data and I would like a GP that gives me an approximate distribution of the samples that will arise in the future, not uncertainty over the posterior.
# Parameters for the solar plant ...
1
vote
0
answers
48
views
Is the GaussianProcessRegressor estimator in scikit-learn able to be parallelized on multi-core?
I do not notice any performance improvements when using GaussianProcessRegressor on a machine which has 8 cores (16 threads), although i do only use physical cores. So I wonder, is the ...
0
votes
2
answers
78
views
`ValueError` in Matrix Multiplication for Gaussian Process Regression Implementation
I'm implementing a Gaussian Process Regression (GPR) model in Python using a Squared Exponential Kernel. However, I'm encountering a ValueError during the matrix multiplication step of the predict ...
-1
votes
1
answer
61
views
Why does mean prediction go flat after more data points are added in Gaussian Process Regressor
Im trying to do a Bayesian optimization in a robot simulator to find optimal Kd and Kp values that fit a desired trajectory (Sinusoidal motion). First I make some random movements of the arm using ...
- 31
0
votes
1
answer
86
views
How to Fix "TypeError: getattr(): attribute name must be string" when multiple optimizers are GridSearched for GaussianProcessRegressor?
Here is my script to predict targets on the final date of a timeseries dataset. I am trying to incorporate a GaussianProcessRegressor model to find the best hyperparameters using GridSearchCV: (Note ...
- 525
0
votes
1
answer
45
views
How to fit a multiple observations to single Gaussian Process
I am trying to fit multiple observations to a single Gaussian Process.
I try to fit the data of two observations (Y) like this:
import numpy as np
from sklearn.gaussian_process import ...
2
votes
0
answers
111
views
Scikit-learn Gaussian Process Regressor returns 2d vector instead of 2d Covariance matrix
I'm fitting data (2d input, 2d output) to a Gaussian Process from Sklearn but when trying to get the covariance matrix I'm getting a 2d vector, not a matrix.
For some examples, it works fine (returns ...
- 473
-1
votes
1
answer
61
views
I can import most of Scikitlearn using miniconda, but cannot get load_digits
I only have miniconda, as i was unable to download anaconda in full, but I am able to import sklearn and hundreds of other programs normally.
However, when I try and import the load_digits part, as ...
1
vote
1
answer
356
views
Are there any resources for coding Gaussian Processes in R?
I am coding one, but the results just don't like as they should. Is there a resource out there that manually performs GP regression on a toy example? Like, do they 1) create the distance function 2) ...
- 77
1
vote
0
answers
25
views
How does a Gaussian process scale with the dimensionality of inputs and outputs?
Let us say we are fitting a GP on a multidimensional dataset.
N - Number of data points [O(N^3), which is known]
D - Dimensionality of each input
M - Dimensionality of each output
Can someone tell ...
- 135
0
votes
1
answer
525
views
What are the different ways to set the length scales of the kernel to be positive and be in the interval range
I'm using the following kernel for my Gaussian process regression with optimizer as Adam. But when i print the model optimized parameters the length scales are not in the interval range i provided.
...
1
vote
0
answers
303
views
How can I force my calculated covariance matrix to be positive definite? _LinAlgError: linalg.cholesky
I am attempting to use Gaussian process regression to estimate a function with two inputs and one output. I know the following about my function:
if x1 == x2, then f(x1, x2) = 0
if f(x1, x2) = y, ...
1
vote
1
answer
6k
views
How to fix ConvergenceWarning in Gaussian process regression in sklearn?
I am trying to use fit a sklearn Gaussian process regressor to my data. The data has periodicity but no mean trend, so I defined a kernel similarly to the tutorial on the Mauna Loa data, without the ...
- 4,417
1
vote
0
answers
548
views
Weighting sample values in scoring methods in scikit-learn
I'm training a Gaussian Process regressor in a mix of categorical and numerical features, and for most categorical features, the amount of data I have is ok, but some categorical features are really ...
2
votes
0
answers
146
views
How do I get n_samples in one call of pm.sample_posterior_predictive() in pymc?
I want to generate multiple samples from one function call of pymc's sample_posterior_predictive(). In the previous version pymc3 there was an argument called 'samples', here is an example https://www....
- 1,855
0
votes
0
answers
143
views
The Covariance Matrix is not positive definite, even with jitter using GPy package in Python
I'm trying to find the optimal variance and lengthscale of a 2D model using Gaussian Process (by using the GPy package in Python). I already found the code to do just that, you can see it below:
# ...
1
vote
1
answer
93
views
Multi-scale Convolutional Neural Network for Image deblurring
I recently reproduce a deep learing model for imaging deblurring, which title is Deep Multi-scale Convolutional Neural Network for Dynamic Scene Deblurring,publised in 2017. The paper mensioned that ...
- 13
0
votes
1
answer
110
views
How to evaluate Gaussian Process Latent Variable Model¶
I am following a tutorial on Gaussian Process Latent Variable Model here is the link https://pyro.ai/examples/gplvm.html
It is a dimension-reduction method.
Now I want to evaluate the model and find ...
- 465
0
votes
0
answers
34
views
Scikit-Learn Decision Tree and Gaussian Process in Leaf
How can I implement a decision tree which performs a Gaussian Process classification in the leafs? I want to do that in the context of an image segmentation.
- 9
0
votes
1
answer
690
views
Define kernel in scikit GaussianProcessRegressor using BayesSearchCV
Question: How do I define the kernel of a Gaussian Process Regressor using BayesSearchCV?
I'm trying to optimize hyperparameters in a gaussian process model using BayesSearchCV from skopt. It seems ...
- 53
0
votes
1
answer
2k
views
Learning multivariate normal covariance matrix using pytorch
I am trying to learn a multivariate normal covariance matrix (Sigma, ∑) using some observations.
The way I went at it is by using pytorch.distributions.MultivariateNormal:
import torch
from torch....
- 166