Show code cell source Hide code cell source

import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
import matplotlib_inline
matplotlib_inline.backend_inline.set_matplotlib_formats('svg')
import seaborn as sns
sns.set_context("paper")
sns.set_style("ticks")

Gaussian Process Regression with Noise

We perform Gaussian process regression with measurement noise. The noise is assumed to be Gaussian with a known variance. The hyper-parameters of the Gaussian process are assumed to be known. See Chapter 3 of [Rasmussen and Williams, 2005] for details.

Example: Gaussian process regression in 1D with fixed hyper-parameters and noise

Let’s generate some synthetic 1D data to work with:

Show code cell source Hide code cell source

np.random.seed(1234)

n = 10
X = np.random.rand(n)
sigma = 0.4
f_true = lambda x: -np.cos(np.pi * x) + np.sin(4. * np.pi * x)
Y = f_true(X) + sigma * np.random.randn(X.shape[0])

fig, ax = plt.subplots()
ax.plot(X, Y, 'kx', markersize=10, markeredgewidth=2)
ax.set_xlabel('$x$')
ax.set_ylabel('$y$')
sns.despine(trim=True);

Again, you may want to install GPyTorch to run this example:

!pip install gpytorch

Let’s make a model like in the previous example. The only difference is that add noise to the likelihood:

import torch
import gpytorch
from gpytorch.kernels import RBFKernel, ScaleKernel

class ExactGP(gpytorch.models.ExactGP):
    def __init__(self, train_x, train_y, likelihood):
        super().__init__(train_x, train_y, likelihood)
        self.mean_module = gpytorch.means.ConstantMean()
        self.covar_module = ScaleKernel(RBFKernel())

    def forward(self, x):
        mean_x = self.mean_module(x)
        covar_x = self.covar_module(x)
        return gpytorch.distributions.MultivariateNormal(mean_x, covar_x)

likelihood = gpytorch.likelihoods.GaussianLikelihood(
    noise_constraint=gpytorch.constraints.GreaterThan(0.0)
)
train_x = torch.from_numpy(X).float()
train_y = torch.from_numpy(Y).float()
model = ExactGP(train_x, train_y, likelihood)

model.mean_module.constant = torch.tensor(0.0)
model.covar_module.outputscale = torch.tensor(1.0)
model.covar_module.base_kernel.lengthscale = torch.tensor(0.1)
model.likelihood.noise = torch.tensor(sigma ** 2)

That’s it. We have now specified the model completely. The posterior GP is completely defined. Where is the posterior mean \(m_n(x)\) and variance \(\sigma_n^2(x)\)? You can get them like this:

model.eval()
likelihood.eval()
# The test points on which we will make predictions
x_star = torch.linspace(0, 1, 100)[:, None]
# The predictive mean and variance for the test points
f_star = model(x_star)
# Now we need to pass through the likelihood - this is different from before
y_star = likelihood(f_star)
# Here is the mean
m_star = y_star.mean
fig, ax = plt.subplots()
ax.plot(X, Y, 'kx', markersize=10, markeredgewidth=2, label='data')
ax.set_xlabel('$x$')
ax.set_ylabel('$y$')
ax.plot(x_star, m_star.detach(), lw=2, label='$m_n(x)$')
plt.legend(loc='best', frameon=False)
sns.despine(trim=True);

Extracting the variance is a bit more involved. Just a tiny bit though. This is because v_star returned by gpm.predict is not exactly \(\sigma_n^2(x)\). It is actually \(\sigma_n^2(x) + \sigma^2\) and not just \(\sigma_n^2(x)\). Here, see it:

fig, ax = plt.subplots()
yv_star = y_star.variance
ax.plot(x_star, yv_star.detach(), lw=2, label='$\sigma_n^2(x)$')
ax.plot(
    x_star,
    model.likelihood.noise.detach() * np.ones(x_star.shape[0]),
    'r--',
    lw=2,
    label='$\sigma^2$'
)

ax.set_ylim(0, 2.1)
plt.legend(loc='best', frameon=False)
ax.set_xlabel('$x$')
ax.set_ylabel('$\sigma_n^2(x)$')
sns.despine(trim=True);

Notice that the variance is small wherever we have an observation. It is not, however, exactly, \(\sigma^2\). It will become exactly \(\sigma^2\) in the limit of many observations. Here \(\sigma^2\) is the aleatory uncertainty. The remaining uncertainty is epistemic uncertainty.

Having the posterior mean and variance, we can derive 95% predictive intervals for \(f(x^*)\) and \(y^*\). For \(f(x^*)\) these are: $\( m_n(\mathbf{x}^*)) - 2\sigma_n(\mathbf{x}^*) \le f(\mathbf{x}^*) \le m_n(\mathbf{x}^*)) + 2\sigma_n(\mathbf{x}^*). \)$ Let’s plot this:

fig, ax = plt.subplots()
ax.set_xlabel('$x$')
ax.set_ylabel('$y$')
v_star = f_star.variance
f_lower = m_star - 2.0 * torch.sqrt(v_star)
f_upper = m_star + 2.0 * torch.sqrt(v_star)
ax.plot(x_star, m_star.detach(), lw=2, label='$m_n(x)$')
ax.fill_between(
    x_star.detach().flatten(),
    f_lower.detach().flatten(),
    f_upper.detach().flatten(),
    alpha=0.5,
    label='$f^*$ 95% CI (epistemic)'
)
ax.plot(X, Y, 'kx', markersize=10, markeredgewidth=2, label='data')
plt.legend(loc='best', frameon=False)
sns.despine(trim=True);

Now, on the same plot, let’s superimpose our predictive error bar about \(y^*\). This is: $\( m_n(\mathbf{x}^*)) - 2\sqrt{\sigma_n^2(\mathbf{x}^*)+\sigma^2}\le f(\mathbf{x}^*) \le m_n(\mathbf{x}^*)) + 2\sqrt{\sigma_n(\mathbf{x}^*) + \sigma^2}. \)$ Let’s use red color for this:

fig, ax = plt.subplots(dpi=100)
ax.set_xlabel('$x$')
ax.set_ylabel('$y$')
y_lower = m_star - 2.0 * torch.sqrt(yv_star)
y_upper = m_star + 2.0 * torch.sqrt(yv_star)
ax.plot(X, Y, 'kx', markersize=10, markeredgewidth=2, label='data')
ax.plot(x_star, m_star.detach(), lw=2, label='$m_n(x)$')
ax.fill_between(
    x_star.detach().flatten(),
    f_lower.detach().flatten(),
    f_upper.detach().flatten(),
    alpha=0.5,
    color=sns.color_palette()[0],
    label='$f^*$ 95% CI (epistemic)'
)
ax.fill_between(
    x_star.detach().flatten(),
    y_lower.detach().flatten(),
    f_lower.detach().flatten(),
    alpha=0.5,
    color=sns.color_palette()[1],
    label='$y^*$ 95% pred.'
)
ax.fill_between(
    x_star.detach().flatten(),
    f_upper.detach().flatten(),
    y_upper.detach().flatten(),
    color=sns.color_palette()[1],
    alpha=0.5,
    label=None
)
plt.legend(loc='best', frameon=False)
sns.despine(trim=True);

Let’s also put the correct function there for comparison:

fig, ax = plt.subplots(dpi=100)
ax.set_xlabel('$x$')
ax.set_ylabel('$y$')
ax.plot(X, Y, 'kx', markersize=10, markeredgewidth=2, label='data')
ax.plot(
    x_star,
    m_star.detach(),
    lw=2,
    label='$m_n(x)$',
    color=sns.color_palette()[0]
)
y_lower = m_star - 2.0 * torch.sqrt(yv_star)
y_upper = m_star + 2.0 * torch.sqrt(yv_star)
ax.fill_between(
    x_star.detach().flatten(),
    y_lower.detach().flatten(),
    f_lower.detach().flatten(),
    color=sns.color_palette()[1],
    alpha=0.5,
    label='$y^*$ 95% CI (aleatory))'
)
ax.fill_between(
    x_star.detach().flatten(),
    f_upper.detach().flatten(),
    y_upper.detach().flatten(),
    color=sns.color_palette()[1],
    alpha=0.5,
    label=None
)
ax.fill_between(
    x_star.detach().flatten(),
    f_lower.detach().flatten(),
    f_upper.detach().flatten(),
    color=sns.color_palette()[0],
    alpha=0.5,
    label='$f^*$ 95% CI (epistemic)'
)
ax.plot(
    x_star,
    f_true(x_star),
    '-.',
    label='True function', 
    color=sns.color_palette()[2]
)
plt.legend(loc='best', frameon=False)
sns.despine(trim=True);

You see that the true function is almost entirely within the blue bounds. It is ok that it is a bit off because these are 95% prediction intervals. About 5% of the function can be off.

Let’s now take some samples from the posterior:

f_post_samples = f_star.sample(sample_shape=torch.Size([10]))

/opt/homebrew/lib/python3.11/site-packages/linear_operator/utils/cholesky.py:40: NumericalWarning: A not p.d., added jitter of 1.0e-06 to the diagonal
  warnings.warn(
/opt/homebrew/lib/python3.11/site-packages/linear_operator/utils/cholesky.py:40: NumericalWarning: A not p.d., added jitter of 1.0e-05 to the diagonal
  warnings.warn(

Let’s plot them along with the data and the truth:

fig, ax = plt.subplots(dpi=100)
ax.plot(x_star, f_post_samples.detach().T, 'r', lw=0.5)
ax.plot([], [], 'r', lw=0.5, label="Posterior samples")
ax.plot(X, Y, 'kx', markersize=10, markeredgewidth=2, label='Observations')
ax.plot(x_star, f_true(x_star), 'm-.', label='True function')
ax.set_xlabel('$x$')
ax.set_ylabel('$y$')
plt.legend(loc="best", frameon=False)
sns.despine(trim=True);

As before, let’s organize the code into a function:

def plot_1d_regression(
    x_star,
    model,
    ax=None,
    f_true=None,
    num_samples=10
):
    """Plot the posterior predictive.
    
    Arguments
    x_start  --  The test points on which to evaluate.
    model    --  The trained model.
    
    Keyword Arguments
    ax          --  An axes object to write on.
    f_true      --  The true function.
    num_samples --  The number of samples.
    """
    f_star = model(x_star)
    m_star = f_star.mean
    v_star = f_star.variance
    y_star = model.likelihood(f_star)
    yv_star = y_star.variance

    f_lower = (
        m_star - 2.0 * torch.sqrt(v_star)
    )
    f_upper = (
        m_star + 2.0 * torch.sqrt(v_star)
    )
    
    y_lower = m_star - 2.0 * torch.sqrt(yv_star)
    y_upper = m_star + 2.0 * torch.sqrt(yv_star)

    if ax is None:
        fig, ax = plt.subplots()
    
    ax.plot(model.train_inputs[0].flatten().detach(),
            model.train_targets.detach(),
            'kx',
            markersize=10,
            markeredgewidth=2,
            label='Observations'
    )

    ax.plot(
        x_star,
        m_star.detach(),
        lw=2,
        label='$m_n(x)$',
        color=sns.color_palette()[0]
    )
    
    ax.fill_between(
        x_star.flatten().detach(),
        f_lower.flatten().detach(),
        f_upper.flatten().detach(),
        alpha=0.5,
        label='$f(\mathbf{x}^*)$ 95% pred.',
        color=sns.color_palette()[0]
    )

    ax.fill_between(
        x_star.detach().flatten(),
        y_lower.detach().flatten(),
        f_lower.detach().flatten(),
        color=sns.color_palette()[1],
        alpha=0.5,
        label='$y^*$ 95% pred.'
    )
    ax.fill_between(
        x_star.detach().flatten(),
        f_upper.detach().flatten(),
        y_upper.detach().flatten(),
        color=sns.color_palette()[1],
        alpha=0.5,
        label=None
    )

    
    if f_true is not None:
        ax.plot(
            x_star,
            f_true(x_star),
            'm-.',
            label='True function'
        )
        
    if num_samples > 0:
        f_post_samples = f_star.sample(
            sample_shape=torch.Size([10])
        )
        ax.plot(
            x_star,
            f_post_samples.T.detach(),
            color="red",
            lw=0.5
        )
        # This is just to add the legend entry
        ax.plot(
            [],
            [],
            color="red",
            lw=0.5,
            label="Posterior samples"
        )
        
    ax.set_xlabel('$x$')
    ax.set_ylabel('$y$')

    plt.legend(loc='best', frameon=False)
    sns.despine(trim=True);

And here is how it works:

plot_1d_regression(x_star, model, f_true=f_true)

/opt/homebrew/lib/python3.11/site-packages/linear_operator/utils/cholesky.py:40: NumericalWarning: A not p.d., added jitter of 1.0e-06 to the diagonal
  warnings.warn(
/opt/homebrew/lib/python3.11/site-packages/linear_operator/utils/cholesky.py:40: NumericalWarning: A not p.d., added jitter of 1.0e-05 to the diagonal
  warnings.warn(

Diagnostics: How do you know if the fit is good?

To objective test the resulting model, we need a validation dataset consisting of the inputs:

\[ \mathbf{x}^v_{1:n^v} = \left(\mathbf{x}^v_1,\dots,\mathbf{x}^v_{n^v}\right), \]

and corresponding observed outputs:

\[ \mathbf{y}^v_{1:n^v} = \left(y^v_1,\dots,y^v_{n^v}\right). \]

We will use this validation dataset to define some diagnostics. Let’s do it directly through the 1D example above. First, we generate some validation data:

n_v = 100
X_v = np.random.rand(n_v)
Y_v = f_true(X_v) + sigma * np.random.randn(n_v)
X_v = torch.from_numpy(X_v).float()
Y_v = torch.from_numpy(Y_v).float()

Point-predictions

Point-predictions only use \(m_n\left(\mathbf{x}^v_i\right)\). Of course, when there is a lot of noise, they are not very useful. But let’s look at what we get anyway. (In the questions section, I will ask you to reduce the noise and repeat).

The simplest thing we can do is to compare \(y^v_i\) to \(m_n\left(\mathbf{x}^v_i\right)\). We start with the mean square error:

\[ \operatorname{MSE} := \frac{1}{n^v}\sum_{i=1}^{n^v}\left[y^v_i-m_n\left(\mathbf{x}^v_i\right)\right]^2. \]

f_v = model(X_v)
y_v = model.likelihood(f_v)
m_v = y_v.mean
v_v = y_v.variance

mse = torch.mean((Y_v - m_v) ** 2)
print(f'MSE = {mse:1.2f}')

MSE = 0.44

This could be more intuitive, though. A somewhat intuitive measure is coefficient of determination, also known as \(R^2\), R squared. It is defined as:

\[ R^2 = 1 - \frac{\sum_{i=1}^{n^v}\left[y_i^v - m_n(\mathbf{x}_i^v)\right]^2}{\sum_{i=1}^{n^v}\left[y_i^v-\bar{y}^v\right]^2}, \]

where \(\bar{y}^v\) is the mean of the observed data:

\[ \bar{y}^v = \frac{1}{n^v}\sum_{i=1}^{n^v}y_i^v. \]

The interpretation of \(R^2\), and take this with a grain of salt, is that it gives the percentage of variance of the data explained by the model. A score of \(R^2=1\), is a perfect fit. In our data, we get:

R2 = 1.0 - torch.sum((Y_v - m_v) ** 2) / torch.sum((Y_v - torch.mean(Y_v)) ** 2)
print(f'R2 = {R2:1.2f}')

R2 = 0.49

Finally, on point-predictions, we can simply plot the predictions vs the observations:

fig, ax = plt.subplots()
y_range = np.linspace(Y_v.min(), Y_v.max(), 50)
ax.plot(y_range, y_range, 'r', lw=2)
ax.plot(Y_v.detach(), m_v.detach(), 'bo')
ax.set_xlabel('Prediction')
ax.set_ylabel('Observation');

Statistical diagnostics

Statistical diagnostics compare the predictive distribution to the distribution of the validation dataset. The way to start are the standardized errors defined by:

\[ e_i = \frac{y_i^v - m_n\left(\mathbf{x}^v_i\right)}{\sigma_n\left(\mathbf{x}^v_i\right)}. \]

If our model is correct, the standardized errors must be distributed as a standard normal \(N(0,1)\) (why?). There are various plots that you can do to test that. First, the histogram of the standardized errors:

import scipy.stats as st

s_v = torch.sqrt(v_v)
e = (Y_v - m_v) / s_v
fig, ax = plt.subplots()
zs = np.linspace(-3.0, 3.0, 100)
ax.plot(zs, st.norm.pdf(zs))
ax.hist(e.detach(), density=True, alpha=0.25)
ax.set_xlabel('Std. error')
sns.despine(trim=True);

Close, but not perfect. Another common plot is this:

fig, ax = plt.subplots()
ax.plot(e.detach(), 'o')
ax.plot(np.arange(e.shape[0]), 2.0 * np.ones(e.shape[0]), 'r--')
ax.plot(np.arange(e.shape[0]), -2.0 * np.ones(e.shape[0]), 'r--')
ax.set_xlabel('$i$')
ax.set_ylabel('$e_i$')
sns.despine(trim=True);

Where the red lines indicate the 95% quantiles of the standard normal. This means that if 5% % of the errors are inside, we are good to go.

Yet another plot yielding the same information is the q-q plot comparing the empirical quantiles of the standardized errors to what they are supposed to be, i.e., to the quantiles of \(N(0,1)\):

fig, ax = plt.subplots(dpi=100)
st.probplot(e.detach().flatten(), dist=st.norm, plot=ax);

Note on Gaussian process diagnostics

For a more detailed description of GP regression diagnostics, please see [Bastos and O'Hagan, 2009].

Questions

• Experiment with different lengthscales for the kernel. What happens to the posterior mean and the 95% predictive error bar as the length scale increases (decreases)?

• Experiment with difference likelihood variances. What happens for huge variances? What happens for tiny variances?

• Experiment with different kernel variances.

• Try some other kernels.

• Experiment with large numbers of training points \(n\). Are the models becoming better according to the metrics we defined above?