diff --git a/README.md b/README.md index 81b3e4d..4ccfa75 100644 --- a/README.md +++ b/README.md @@ -117,7 +117,7 @@ print(problem.names, sampler.results.logz[-1]) - [`docs/recipes.md`](docs/recipes.md): every supported use case with its spelling and expected behaviour. Each recipe is a test under `test/recipes/`. -- [`examples/`](examples/): nine notebooks, the tutorials for the recipes, +- [`examples/`](examples/): the notebooks, the tutorials for the recipes, from a line to an error-model comparison on real α + ⁴⁴Ca data and a hierarchical calibration. - [`docs/design.md`](docs/design.md): the maintainer's description of the diff --git a/docs/api.rst b/docs/api.rst index 98d5de3..0bdfa43 100644 --- a/docs/api.rst +++ b/docs/api.rst @@ -52,6 +52,7 @@ Covariance terms rxmc.terms.proportional_error rxmc.terms.systematic rxmc.terms.kernel + rxmc.terms.kernel_param_names rxmc.terms.ones rxmc.terms.ym rxmc.terms.averaging diff --git a/docs/design.md b/docs/design.md index 6fbf046..fa6d561 100644 --- a/docs/design.md +++ b/docs/design.md @@ -243,6 +243,7 @@ proportional_error(parameter, averaging=False, log=True, on=None) systematic(parameter, basis, log=True, basis_params=(), on=None, coords=None) kernel(kernel, coords=None, amplitude=None, amplitude_params=(), jitter=1e-10, prefix="discrepancy", params=None, on=None) -> KernelTerm +kernel_param_names(kernel) -> list[str] # the keys of a params= mapping # bases and amplitudes: ones, ym, averaging, x_basis(scale), exp_growth(scale, base=ones), # constant_amplitude, exp_growth_amplitude(scale) ``` @@ -253,7 +254,11 @@ reads `c.ym`, never `c.y`. `kernel` derives one `Parameter` per free hyperparameter element in sklearn's log-theta space, bounded by the log of the kernel's bounds so it compiles with a uniform prior there; `params=` passes the objects instead, which is also how hyperparameters are shared -between per-comparison kernels. Two kernel terms with derived names and +between per-comparison kernels. `params=` is a mapping keyed by +`kernel_param_names(kernel)` (sklearn's `theta` order is alphabetical by +attribute name); a positional list is accepted only for a kernel with one +free hyperparameter, because with two or more a misread order would swap +priors silently. Two kernel terms with derived names and one prefix fail compile on the duplicate name. Hierarchy is sharing plus `meta`: a hyperparameter shared by several @@ -652,7 +657,7 @@ Two tiers, chosen per assertion. Three index tests keep the documents honest: `test_recipes_index.py` (one file per `## NN.` heading, each quoting its recipe; the recipe-18 legend equal to the tests' legend), `test_notebooks_index.py` (each -notebook cites the recipes of the design's table, all nine exist), and +notebook cites the recipes of the design's table, and every one exists), and `test_regression.py` (the 0.x pins). `test/helpers.py` holds the dense references and `STUDY_LEGEND`, the labelled error-model forms of recipe 18 built against hand-written matrices. @@ -670,6 +675,7 @@ a time unless noted. | `sharing_error_models` | 5 | dynesty | two experiments with opposite normalisation defects: sharing a parameter, a mode per dataset, one mode spanning both, and the assembled covariance seen directly; one panel per error model | 102 s | | `normalization_and_covariance_structure` | 3, 4, 6, 27 | dynesty | six treatments of four experiments' normalisations on a cubic with a free, non-zero constant term, one of them badly mis-quoted and alone in its range; Peelle's puzzle in the two-point case it was found in; a gallery of covariance structures from `matrix(theta)` | 1003 s | | `gp_discrepancy` | 7 | emcee, dynesty | mean-zero discrepancies with amplitudes growing in x: four rungs on a toy line, three on n+⁴⁰Ca missing its surface absorption; the three predictive objects (model plus discrepancy, plus experimental, and the conditioned regression it does not use), with the equations | 483 s | +| `spectral_mixture_discrepancy` | 7, 33, 40 | MAP, emcee | a defect with two scales on a line, a three-component spectral-mixture discrepancy written as a user sklearn kernel and keyed `params=`; periodogram-seeded multistart MAP; each component backed out with `gp_predictive_draws(conditioned=True)`; the identifiability of θ, with the L² projection θ* as the only recoverable target; Plumlee's orthogonal kernel as a second user kernel; a chain against each Laplace approximation, which agrees for Plumlee and is several times too narrow and misplaced for Kennedy–O'Hagan | 930 s | | `robust_likelihoods` | 9, 12, 39 | emcee | Student-t versus Gaussian on three gross outliers; the iterative rejection loop, including the round that over-rejects and recovers; tempering a correctly specified 200-point line, with the prior on the corner plot and the predictive coverage | 80 s | | `error_scale_and_usu` | 34 | emcee | a global scale on the reported errors under both likelihoods; a USU offset on the technique we suspect | 90 s | | `local_optical_model_calibration` | 10, 14, 15, 16, 17, 18, 19, 21, 40 | dynesty | EXFOR O1199007, p + ⁴⁰Ca at 35 MeV, which quotes no systematics, read from a plain dict: the unit contract; log-space residuals with the jitr α+Ca error model (reported statistics plus inferred point-to-point and common relative errors, as custom terms) against linear residuals with the reported errors alone, compared by evidence with the Jacobian, by predictives for the potential alone (on a grid) and with every error term (at the measured angles), and by three coverage curves each; the singular guard | 929 s | diff --git a/docs/examples.rst b/docs/examples.rst index 68ae886..55251ff 100644 --- a/docs/examples.rst +++ b/docs/examples.rst @@ -27,6 +27,7 @@ Beyond the Gaussian :maxdepth: 1 examples/gp_discrepancy.ipynb + examples/spectral_mixture_discrepancy.ipynb examples/robust_likelihoods.ipynb examples/error_scale_and_usu.ipynb diff --git a/docs/groundup_design.md b/docs/groundup_design.md index f2a95b7..7e999fa 100644 --- a/docs/groundup_design.md +++ b/docs/groundup_design.md @@ -314,7 +314,8 @@ kernel(kernel, coords=None, amplitude=None, amplitude_params=(), jitter=1e-10, # KernelTerm(Term) adds kernel, n_kernel, amplitude, jitter so predictive.gp_predictive_draws # can condition the discrepancy from the term alone. A derived hyperparameter is # bounded by the log of the kernel's bounds (a uniform prior in log-theta). -# params=: the hyperparameter Parameter objects, one per free element in kernel.theta order; +# params=: the hyperparameter Parameter objects, keyed by name (kernel_param_names; positional +# only for a single free hyperparameter; superseded detail: see design.md); # None derives fresh ones named f"{prefix}_{name}". Pass the same objects to share # hyperparameters between per-block kernels. Two kernel terms with derived # names and the same prefix fail compile on the duplicate name; the error says diff --git a/docs/recipes.md b/docs/recipes.md index d7332de..2b189e2 100644 --- a/docs/recipes.md +++ b/docs/recipes.md @@ -638,6 +638,12 @@ Expected behaviour: - Without `params=[ell]`, each `kernel` call would derive its own length scale, one per dataset. That is also a legitimate model; it is just not this one. +- `params=[ell]` is positional because Matern here has one free + hyperparameter. A kernel with several takes a mapping keyed by + `T.kernel_param_names(k)`, e.g. for `ExpSineSquared` + `params={"length_scale": ell, "periodicity": per}`; + a positional list for it is refused, since sklearn orders hyperparameters + alphabetically and a misread order would swap priors silently. - The comparisons stay uncorrelated, so the constraint is on the structured path. diff --git a/examples/normalization_and_covariance_structure.ipynb b/examples/normalization_and_covariance_structure.ipynb index a8300e7..b4927a0 100644 --- a/examples/normalization_and_covariance_structure.ipynb +++ b/examples/normalization_and_covariance_structure.ipynb @@ -60,7 +60,7 @@ "id": "240f15bb-1321-402c-915b-ef269093abfd", "metadata": {}, "source": [ - "### Generate synthetic data" + "## Generate synthetic data" ] }, { diff --git a/examples/spectral_mixture_discrepancy.ipynb b/examples/spectral_mixture_discrepancy.ipynb new file mode 100644 index 0000000..e38df6c --- /dev/null +++ b/examples/spectral_mixture_discrepancy.ipynb @@ -0,0 +1,1023 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "id": "15102cda", + "metadata": {}, + "source": [ + "# A discrepancy with several length scales: the spectral-mixture kernel\n", + "\n", + "The `gp_discrepancy` notebook gave the model's error one kernel with one length scale. Real model defects are rarely that tidy. A missing mechanism can leave a slow drift *and* a fast oscillation in the same residuals, and a single stationary kernel has to compromise between them: too smooth for one, too rough for the other.\n", + "\n", + "We would rather not guess the shape of the covariance. The [spectral-mixture kernel](https://proceedings.mlr.press/v28/wilson13.html) of Wilson & Adams (2013) learns it instead. By [Bochner's theorem](https://en.wikipedia.org/wiki/Bochner%27s_theorem) a stationary kernel is the Fourier transform of a spectral density, and any spectral density can be approximated by a mixture of Gaussians. Each Gaussian in frequency is one *component* of the discrepancy, with its own amplitude, characteristic frequency and bandwidth. If the data contain two scales, the posterior should switch two components on.\n", + "\n", + "That is the first half of this notebook: detect the scales, then back out the component each one contributes to the residuals. The second half is about what the discrepancy does to the physics parameters. A discrepancy that can absorb anything can absorb part of the model, and then $\\theta$ is no longer identified ([Brynjarsdóttir & O'Hagan, 2014](https://doi.org/10.1088/0266-5611/30/11/114007)). We will see that happen, and then use the orthogonality condition of [Plumlee (2017)](https://doi.org/10.1080/01621459.2017.1329149) to say precisely which $\\theta$ the calibration *can* recover.\n", + "\n", + "We fit by maximum a posteriori with a Laplace approximation, and check both against a chain at the end. The check matters: for one of the two discrepancy models it overturns the Laplace answer.\n", + "\n", + "Recipes: 7, 33, 40" + ] + }, + { + "cell_type": "code", + "execution_count": 1, + "id": "b49d1d32", + "metadata": {}, + "outputs": [], + "source": [ + "from functools import lru_cache\n", + "\n", + "import emcee\n", + "import matplotlib.pyplot as plt\n", + "import numpy as np\n", + "import plotstyle\n", + "from matplotlib.patches import Ellipse\n", + "from scipy import stats\n", + "from scipy.optimize import minimize\n", + "from scipy.signal import lombscargle\n", + "from sklearn.gaussian_process.kernels import Hyperparameter, Kernel\n", + "\n", + "import rxmc as rx\n", + "from rxmc import terms as T\n", + "\n", + "plotstyle.use()" + ] + }, + { + "cell_type": "markdown", + "id": "1b56e108", + "metadata": {}, + "source": [ + "## A model, and a defect with two scales\n", + "\n", + "The model is a straight line, $\\eta(x;\\theta) = \\theta_0 + \\theta_1 x$, on $x \\in [0, 10]$. Reality adds two independent defects to it:\n", + "\n", + "- a **slow drift**, drawn from a squared-exponential GP with amplitude 0.3 and length scale 2;\n", + "- a **fast oscillation** of period 0.7 under a slowly varying envelope (amplitude 0.15, envelope length 3).\n", + "\n", + "The data are 120 points at random $x$ with a known, honestly reported noise $\\sigma = 0.05$. We draw each defect jointly on the measured points and on a fine grid, so that we can plot it and integrate it later." + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "id": "22c04f9e", + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "Dataset('toy', n=120)" + ] + }, + "execution_count": 2, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "rng = np.random.default_rng(1)\n", + "n, sigma, L = 120, 0.05, 10.0\n", + "x = np.sort(rng.uniform(0.0, L, n))\n", + "x_fine = np.linspace(0.0, L, 401)\n", + "theta_true = np.array([2.0, 0.5])\n", + "\n", + "\n", + "def h(z):\n", + " \"\"\"d eta / d theta: the line's basis (1, x).\"\"\"\n", + " return np.column_stack([np.ones_like(z), z])\n", + "\n", + "\n", + "z = np.concatenate([x, x_fine])\n", + "tau = z[:, None] - z[None, :]\n", + "K_drift = 0.30**2 * np.exp(-0.5 * tau**2 / 2.0**2)\n", + "K_wiggle = 0.15**2 * np.exp(-0.5 * tau**2 / 3.0**2) * np.cos(2 * np.pi * tau / 0.7)\n", + "jitter = 1e-8 * np.eye(z.size)\n", + "drift = np.linalg.cholesky(K_drift + jitter) @ rng.standard_normal(z.size)\n", + "wiggle = np.linalg.cholesky(K_wiggle + jitter) @ rng.standard_normal(z.size)\n", + "delta_true = drift + wiggle\n", + "\n", + "y = h(x) @ theta_true + delta_true[:n] + sigma * rng.standard_normal(n)\n", + "data = rx.Dataset(x, y, np.full(n, sigma), label=\"toy\")\n", + "data" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "id": "737421ab", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, (ax0, ax1) = plt.subplots(2, 1, figsize=(6.4, 5.6), sharex=True)\n", + "ax0.errorbar(x, y, sigma, fmt=\"o\", color=\"k\", ms=2.5, label=\"data\")\n", + "ax0.plot(\n", + " x_fine,\n", + " h(x_fine) @ theta_true,\n", + " \"--\",\n", + " color=plotstyle.COLOURS[1],\n", + " label=r\"$\\eta(x;\\theta_{\\rm true})$\",\n", + ")\n", + "ax0.set(ylabel=\"$y$\")\n", + "ax0.legend()\n", + "ax1.plot(x_fine, drift[n:], color=plotstyle.COLOURS[0], label=\"slow drift\")\n", + "ax1.plot(x_fine, wiggle[n:], color=plotstyle.COLOURS[2], label=\"fast oscillation\")\n", + "ax1.set(xlabel=\"$x$\", ylabel=r\"$\\delta(x)$\")\n", + "ax1.legend(ncol=2)\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "269930ba", + "metadata": {}, + "source": [ + "## One spectral-mixture component as an sklearn kernel\n", + "\n", + "A single component with spectral mean $\\mu$ and spectral standard deviation $s$ has the covariance\n", + "\n", + "$$k(\\tau) = \\exp\\!\\left(-2\\pi^2 s^2 \\tau^2\\right) \\cos(2\\pi\\mu\\tau),$$\n", + "\n", + "the Fourier transform of a Gaussian at $\\pm\\mu$ with width $s$. $1/\\mu$ is the period of the oscillation and $1/(2\\pi s)$ the length of its envelope; $\\mu = 0$ is the familiar squared-exponential kernel.\n", + "\n", + "`T.kernel` accepts any object with sklearn's kernel interface, so a new kernel is a small `Kernel` subclass. The amplitude is left out: rxmc supplies it with `T.constant_amplitude`, exactly as for the built-in kernels. The two hyperparameters, `mu` and `s`, are declared with `Hyperparameter`, and sklearn samples them in log space." + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "id": "26c0c6ab", + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "['mu', 's']" + ] + }, + "execution_count": 4, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "class SpectralComponent(Kernel):\n", + " r\"\"\"One spectral-mixture component, :math:`e^{-2\\pi^2 s^2\\tau^2}\\cos(2\\pi\\mu\\tau)`.\"\"\"\n", + "\n", + " def __init__(self, mu=0.5, s=0.3, mu_bounds=(1e-3, 1e2), s_bounds=(1e-3, 1e2)):\n", + " self.mu, self.s = mu, s\n", + " self.mu_bounds, self.s_bounds = mu_bounds, s_bounds\n", + "\n", + " @property\n", + " def hyperparameter_mu(self):\n", + " return Hyperparameter(\"mu\", \"numeric\", self.mu_bounds)\n", + "\n", + " @property\n", + " def hyperparameter_s(self):\n", + " return Hyperparameter(\"s\", \"numeric\", self.s_bounds)\n", + "\n", + " def _k(self, X, Y):\n", + " tau = X[:, :1] - Y[:, :1].T\n", + " return np.exp(-2 * np.pi**2 * self.s**2 * tau**2) * np.cos(\n", + " 2 * np.pi * self.mu * tau\n", + " )\n", + "\n", + " def __call__(self, X, Y=None, eval_gradient=False):\n", + " if eval_gradient:\n", + " raise NotImplementedError(\"rxmc never asks a kernel for gradients\")\n", + " X = np.atleast_2d(X)\n", + " return self._k(X, X if Y is None else np.atleast_2d(Y))\n", + "\n", + " def diag(self, X):\n", + " return np.ones(np.atleast_2d(X).shape[0])\n", + "\n", + " def is_stationary(self):\n", + " return True\n", + "\n", + "\n", + "T.kernel_param_names(SpectralComponent())" + ] + }, + { + "cell_type": "markdown", + "id": "4d411ef4", + "metadata": {}, + "source": [ + "sklearn orders a kernel's hyperparameters alphabetically by attribute name, not in the order of `__init__`. Here that happens to be `mu, s`; renaming `s` to `bandwidth` would reverse it. That is why `T.kernel` takes a kernel with more than one free hyperparameter's parameters as a mapping keyed by these names, and refuses a positional list.\n", + "\n", + "## Priors, bounds, and the problem\n", + "\n", + "We give the discrepancy $Q = 3$ components, one more than the truth has, to see whether the surplus switches off. Each component has three parameters, all sampled as logarithms:\n", + "\n", + "- **the amplitude** $a_q$, with a half-Cauchy prior of scale 0.3, which puts mass near zero (a component can turn itself off) but has a heavy tail (it is not forbidden to matter). The half-Cauchy on $a$ is exactly a hyperbolic-secant distribution on $\\log a$, so it is a plain scipy prior: if $a \\sim {\\rm HalfCauchy}(0, c)$ then $\\log a$ has density $\\operatorname{sech}(\\log a - \\log c)/\\pi$;\n", + "- **the frequency** $\\mu_q$ and **bandwidth** $s_q$, with wide log-normal priors.\n", + "\n", + "The bounds are not decoration. The design can only resolve frequencies between roughly one cycle in five domains and the [Nyquist frequency](https://en.wikipedia.org/wiki/Nyquist_frequency) of the mean point spacing, and envelopes no shorter than about the spacing. A component outside that band is indistinguishable from the linear model (too slow) or from the noise (too fast), and its hyperparameters would wander there without bound." + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "id": "fb380559", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Problem(ndim=11, constraints=1)\n", + "['theta0', 'theta1', 'log_mu0', 'log_s0', 'log_a0', 'log_mu1', 'log_s1', 'log_a1', 'log_mu2', 'log_s2', 'log_a2']\n" + ] + } + ], + "source": [ + "Q = 3\n", + "log_mu_bounds = (np.log(0.2 / L), np.log(n / (2 * L))) # period 50 ... Nyquist\n", + "log_s_bounds = (np.log(0.005), np.log(2.0))\n", + "log_a_bounds = (np.log(1e-3), np.log(10.0))\n", + "\n", + "theta0 = rx.Parameter(\"theta0\", prior=stats.norm(0.0, 10.0), latex=r\"\\theta_0\")\n", + "theta1 = rx.Parameter(\"theta1\", prior=stats.norm(0.0, 10.0), latex=r\"\\theta_1\")\n", + "line = rx.Model(lambda x, theta0, theta1: theta0 + theta1 * x, [theta0, theta1])\n", + "comp = rx.Comparison(data, line)\n", + "\n", + "\n", + "def components(make_kernel):\n", + " \"\"\"Q kernel terms on `comp`, each with its own amplitude, frequency, bandwidth.\"\"\"\n", + " terms = []\n", + " for q in range(Q):\n", + " log_mu = rx.Parameter(\n", + " f\"log_mu{q}\", bounds=log_mu_bounds, prior=stats.norm(np.log(0.5), 2.0)\n", + " )\n", + " log_s = rx.Parameter(\n", + " f\"log_s{q}\", bounds=log_s_bounds, prior=stats.norm(np.log(0.3), 2.0)\n", + " )\n", + " log_a = rx.Parameter(\n", + " f\"log_a{q}\", bounds=log_a_bounds, prior=stats.hypsecant(loc=np.log(0.3))\n", + " )\n", + " terms.append(\n", + " T.kernel(\n", + " make_kernel(),\n", + " params={\"mu\": log_mu, \"s\": log_s},\n", + " amplitude=T.constant_amplitude,\n", + " amplitude_params=(log_a,),\n", + " on=comp,\n", + " jitter=1e-8,\n", + " )\n", + " )\n", + " return terms\n", + "\n", + "\n", + "koh_terms = components(SpectralComponent)\n", + "koh = rx.Problem([rx.Constraint([comp], terms=koh_terms)])\n", + "print(koh)\n", + "print(koh.names)" + ] + }, + { + "cell_type": "markdown", + "id": "1ad771dc", + "metadata": {}, + "source": [ + "The constraint keeps its default `statistical=True`, so the reported $\\sigma$ sits on the diagonal and the three components are added to it. This is the [Kennedy & O'Hagan (2001)](https://doi.org/10.1111/1467-9868.00294) model: $y = \\eta(x;\\theta) + \\delta(x) + \\varepsilon$ with $\\delta = \\sum_q \\delta_q$ mean-zero, integrated out of the likelihood.\n", + "\n", + "## Finding the posterior mode\n", + "\n", + "The spectral-mixture likelihood is notoriously multimodal in the frequencies, so random restarts waste most of their effort. A better start reads the frequencies off the data. We fit the line by least squares and compute the [Lomb–Scargle periodogram](https://en.wikipedia.org/wiki/Least-squares_spectral_analysis) of the residuals, which handles unevenly spaced points, and put one component on each of its $Q$ largest well-separated peaks." + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "id": "dda29080", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "peak at period 0.71, amplitude ~ 0.200\n", + "peak at period 8.17, amplitude ~ 0.090\n", + "peak at period 1.09, amplitude ~ 0.080\n" + ] + } + ], + "source": [ + "def periodogram_peaks(Q):\n", + " \"\"\"(frequency, amplitude) of the Q largest separated peaks of the OLS residual.\"\"\"\n", + " r = y - h(x) @ np.linalg.lstsq(h(x), y, rcond=None)[0]\n", + " f = np.exp(np.linspace(*log_mu_bounds, 2000))\n", + " P = lombscargle(x, r, 2 * np.pi * f, precenter=True)\n", + " peaks = [i for i in range(1, f.size - 1) if P[i - 1] <= P[i] >= P[i + 1]]\n", + " seeds = []\n", + " for i in sorted(peaks, key=lambda i: -P[i]):\n", + " if all(abs(np.log(f[i] / g)) > 0.3 for g, _ in seeds):\n", + " seeds.append((f[i], np.sqrt(2 * P[i] / n)))\n", + " if len(seeds) == Q:\n", + " return seeds\n", + " return seeds\n", + "\n", + "\n", + "seeds = periodogram_peaks(Q)\n", + "for f, a in seeds:\n", + " print(f\"peak at period {1 / f:6.2f}, amplitude ~ {a:.3f}\")" + ] + }, + { + "cell_type": "markdown", + "id": "bcff274e", + "metadata": {}, + "source": [ + "The two real scales are the two largest peaks: period about 0.71, and a slow mode at period about 8 that is the drift seen through a ten-unit window. The third peak, at period 1.09, is the surplus component's starting point.\n", + "\n", + "From there we minimise $-\\log p(\\theta, \\phi \\mid y)$ with L-BFGS-B inside `problem.bounds` (recipe 33). The first start sits exactly on the peaks, with an envelope one spectral resolution wide; the others jitter it by a factor of about $e$." + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "id": "d6e45859", + "metadata": {}, + "outputs": [], + "source": [ + "def fit_map(problem, n_starts=6, seed=0):\n", + " \"\"\"Multistart L-BFGS-B from the periodogram seeds, inside problem.bounds.\"\"\"\n", + " r = np.random.default_rng(seed)\n", + " names, bounds = problem.names, problem.bounds\n", + " start = np.empty(problem.ndim)\n", + " start[problem.columns(line.params)] = np.linalg.lstsq(h(x), y, rcond=None)[0]\n", + " best = None\n", + " for k in range(n_starts):\n", + " t0 = start.copy()\n", + " for q in range(Q):\n", + " f, a = seeds[q % len(seeds)]\n", + " jit = np.zeros(3) if k == 0 else r.normal(0.0, 1.0, 3)\n", + " t0[names.index(f\"log_mu{q}\")] = np.log(f) + jit[0]\n", + " t0[names.index(f\"log_s{q}\")] = np.log(1.0 / L) + jit[1]\n", + " t0[names.index(f\"log_a{q}\")] = np.log(a) + jit[2]\n", + " t0 = np.clip(t0, bounds[:, 0] + 1e-6, bounds[:, 1] - 1e-6)\n", + " res = minimize(\n", + " lambda t: -problem.log_posterior(t), t0, method=\"L-BFGS-B\", bounds=bounds\n", + " )\n", + " if best is None or res.fun < best.fun:\n", + " best = res\n", + " return best.x\n", + "\n", + "\n", + "def describe(problem, terms, t):\n", + " \"\"\"One line per component: amplitude, period, envelope length.\"\"\"\n", + " for q, term in enumerate(terms):\n", + " log_mu, log_s, log_a = t[problem.columns(term.params)]\n", + " print(\n", + " f\"component {q}: amplitude {np.exp(log_a):.3f} period \"\n", + " f\"{np.exp(-log_mu):6.2f} envelope {1 / (2 * np.pi * np.exp(log_s)):6.2f}\"\n", + " )" + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "id": "5cb284f7", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "component 0: amplitude 0.178 period 0.72 envelope 9.24\n", + "component 1: amplitude 0.110 period 7.21 envelope 14.94\n", + "component 2: amplitude 0.161 period 0.68 envelope 5.93\n", + "theta_hat = [2.147 0.452]\n", + "CPU times: user 3min 6s, sys: 48.4 ms, total: 3min 6s\n", + "Wall time: 3min 8s\n" + ] + } + ], + "source": [ + "%%time\n", + "koh_map = fit_map(koh)\n", + "describe(koh, koh_terms, koh_map)\n", + "print(\"theta_hat =\", koh_map[koh.columns(line.params)].round(3))" + ] + }, + { + "cell_type": "markdown", + "id": "826b4cd6", + "metadata": {}, + "source": [ + "## What the components found\n", + "\n", + "The learned spectral density is the sum of the components' Gaussians, weighted by $a_q^2$. The truth has one of its own: the drift is a Gaussian at zero frequency with width $1/(2\\pi \\cdot 2)$, and the oscillation a Gaussian at $1/0.7$ with width $1/(2\\pi\\cdot 3)$." + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "id": "1725e86b", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "def spectrum(f, amp2, mu, s):\n", + " \"\"\"A symmetrised spectral Gaussian, on positive frequencies.\"\"\"\n", + " g = stats.norm(mu, s)\n", + " return amp2 * (g.pdf(f) + g.pdf(-f))\n", + "\n", + "\n", + "f_grid = np.linspace(0.0, 2.5, 1500)\n", + "fig, ax = plt.subplots()\n", + "ax.fill_between(\n", + " f_grid,\n", + " spectrum(f_grid, 0.30**2, 0.0, 1 / (2 * np.pi * 2.0))\n", + " + spectrum(f_grid, 0.15**2, 1 / 0.7, 1 / (2 * np.pi * 3.0)),\n", + " color=\"0.85\",\n", + " label=\"truth\",\n", + ")\n", + "for q, term in enumerate(koh_terms):\n", + " log_mu, log_s, log_a = koh_map[koh.columns(term.params)]\n", + " ax.plot(\n", + " f_grid,\n", + " spectrum(f_grid, np.exp(2 * log_a), np.exp(log_mu), np.exp(log_s)),\n", + " label=f\"component {q}\",\n", + " )\n", + "ax.set(\n", + " xlabel=\"frequency (cycles per unit $x$)\",\n", + " ylabel=\"spectral density\",\n", + " yscale=\"log\",\n", + " ylim=(1e-4, 10),\n", + ")\n", + "ax.legend()\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "20bff9d4", + "metadata": {}, + "source": [ + "Both true scales are found, but not one component each. Component 1 carries the drift, as a narrow peak near frequency 0.14 rather than a broad peak at zero: over a window of five drift lengths the data contain about a cycle and a half of the drift, and a slow sinusoid fits that as well as a squared exponential does. The spectral *parameters* of a component whose scale is comparable to the window are not interpretable; only its existence is.\n", + "\n", + "The surplus component did not switch off. Components 0 and 2 share the oscillation between them, at periods 0.72 and 0.68, both with large amplitudes. Splitting one real mode between two components costs almost nothing under a half-Cauchy amplitude prior, so the count of *distinct spectral modes* (two) is the meaningful number here, not the count of active components (three).\n", + "\n", + "### Backing out each component\n", + "\n", + "Because the discrepancy is a sum of independent GPs, each component has its own posterior given the data, with every other term of the covariance acting as noise:\n", + "\n", + "$$\\mathbb E[\\delta_q \\mid y] = K_q (K + \\Sigma_{\\rm exp})^{-1}(y - \\eta(\\hat\\theta)).$$\n", + "\n", + "That is what `gp_predictive_draws(..., terms=[term], conditioned=True)` computes, on any grid we bind the model to (recipe 40): its default conditioning noise is everything in the constraint except the chosen term. The bands below are conditional on the modal hyperparameters; we subtract the model's own prediction to leave $\\delta_q$." + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "id": "1c0b6e52", + "metadata": {}, + "outputs": [], + "source": [ + "on_fine = line.bind(x_fine)\n", + "\n", + "\n", + "def component_bands(problem, terms, t):\n", + " \"\"\"(16, 50, 84) percentile bands of each delta_q on x_fine, at hyperparameters t.\"\"\"\n", + " eta = on_fine(*t[problem.columns(line.params)])\n", + " return [\n", + " rx.predictive.gp_predictive_draws(\n", + " problem,\n", + " term,\n", + " on_fine,\n", + " x_fine,\n", + " t[None, :],\n", + " terms=[term],\n", + " conditioned=True,\n", + " levels=(16, 50, 84),\n", + " n_rep=400,\n", + " rng=q,\n", + " )\n", + " - eta\n", + " for q, term in enumerate(terms)\n", + " ]\n", + "\n", + "\n", + "koh_bands = component_bands(koh, koh_terms, koh_map)" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "id": "fb0d6967", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "def plot_components(bands, t, problem, terms, title):\n", + " fig, axes = plt.subplots(Q, 1, figsize=(6.4, 6.4), sharex=True, sharey=True)\n", + " for q, (ax, band) in enumerate(zip(axes, bands)):\n", + " period = np.exp(-t[problem.columns(terms[q].params)][0])\n", + " plotstyle.band(\n", + " ax,\n", + " x_fine,\n", + " band[0],\n", + " band[2],\n", + " color=plotstyle.COLOURS[q],\n", + " label=f\"component {q} (period {period:.2f}), 68 %\",\n", + " )\n", + " ax.plot(x_fine, drift[n:], \"--\", color=\"0.35\", lw=1, label=\"true drift\")\n", + " ax.plot(x_fine, wiggle[n:], \":\", color=\"0.35\", lw=1, label=\"true oscillation\")\n", + " ax.legend(fontsize=7, ncol=3, loc=\"lower left\")\n", + " ax.set(ylabel=rf\"$\\delta_{q}$\")\n", + " axes[0].set_title(title)\n", + " axes[-1].set(xlabel=\"$x$\")\n", + " plt.show()\n", + "\n", + "\n", + "plot_components(koh_bands, koh_map, koh, koh_terms, \"Kennedy-O'Hagan discrepancy\")" + ] + }, + { + "cell_type": "markdown", + "id": "0f0027bd", + "metadata": {}, + "source": [ + "Components 0 and 2 each carry part of the oscillation; their sum follows it. Component 1 follows the drift's shape but not its level: it sits offset and tilted against the true drift. That is not a failure of the component. It is the subject of the rest of this notebook.\n", + "\n", + "## Which $\\theta$ did we calibrate?" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "id": "1817a93b", + "metadata": {}, + "outputs": [], + "source": [ + "def laplace(problem, t, step=1e-4, tol=1e-3):\n", + " \"\"\"Inverse Hessian of -log posterior over the coordinates off the bounds.\n", + "\n", + " A hyperparameter that converged onto a bound has no Gaussian neighbourhood;\n", + " it is held fixed, and the covariance is conditional on it.\n", + " \"\"\"\n", + " lo, hi = problem.bounds[:, 0], problem.bounds[:, 1]\n", + " free = np.flatnonzero(np.minimum(t - lo, hi - t) > tol)\n", + " f = lambda v: -problem.log_posterior(v)\n", + " E = np.zeros((free.size, t.size))\n", + " E[np.arange(free.size), free] = step\n", + " H = np.array(\n", + " [\n", + " [f(t + a + b) - f(t + a - b) - f(t - a + b) + f(t - a - b) for b in E]\n", + " for a in E\n", + " ]\n", + " ) / (4 * step**2)\n", + " cov = np.full((t.size, t.size), np.nan)\n", + " cov[np.ix_(free, free)] = np.linalg.inv(0.5 * (H + H.T))\n", + " pinned = [problem.names[i] for i in range(t.size) if i not in free]\n", + " return cov, pinned\n", + "\n", + "\n", + "def report(problem, t, label):\n", + " cov, pinned = laplace(problem, t)\n", + " cols = problem.columns(line.params)\n", + " th, C = t[cols], cov[np.ix_(cols, cols)]\n", + " sd = np.sqrt(np.diag(C))\n", + " print(f\"{label}: theta_hat = {th.round(3)} +/- {sd.round(3)}\")\n", + " if pinned:\n", + " print(\" on a bound, held fixed:\", pinned)\n", + " for name, ref in [(\"theta_true\", theta_true), (\"theta*\", theta_star)]:\n", + " d = th - ref\n", + " print(\n", + " f\" Mahalanobis distance to {name:10s}: {np.sqrt(d @ np.linalg.solve(C, d)):6.2f}\"\n", + " )\n", + " return th, C" + ] + }, + { + "cell_type": "markdown", + "id": "34a5b8d3", + "metadata": {}, + "source": [ + "The truth is $\\theta = (2, 0.5)$, and the fit will not land there. Any discrepancy can be split into a part the line can represent and a part it cannot. The likelihood sees only $\\eta + \\delta$, so it has no way to assign the first part to $\\delta$ rather than to $\\theta$. What the data *can* pin down is the projection of reality onto the model,\n", + "\n", + "$$\\theta^\\ast = \\arg\\min_\\theta \\int_0^{10} \\big(\\zeta(x) - \\eta(x;\\theta)\\big)^2\\, dx, \\qquad \\zeta = \\eta(\\cdot\\,;\\theta_{\\rm true}) + \\delta,$$\n", + "\n", + "which we can compute here because we know $\\delta$ on the fine grid." + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "id": "587afe18", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "theta_true = [2. 0.5] theta* = [2.203 0.445]\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Kennedy-O'Hagan: theta_hat = [2.147 0.452] +/- [0.023 0.004]\n", + " Mahalanobis distance to theta_true: 18.74\n", + " Mahalanobis distance to theta* : 4.41\n" + ] + } + ], + "source": [ + "w = np.full(x_fine.size, x_fine[1] - x_fine[0])\n", + "w[[0, -1]] *= 0.5 # trapezoid\n", + "G = h(x_fine).T @ (w[:, None] * h(x_fine))\n", + "theta_star = theta_true + np.linalg.solve(G, h(x_fine).T @ (w * delta_true[n:]))\n", + "print(\"theta_true =\", theta_true, \" theta* =\", theta_star.round(3))\n", + "koh_theta, koh_cov = report(koh, koh_map, \"Kennedy-O'Hagan\")" + ] + }, + { + "cell_type": "markdown", + "id": "52df1275", + "metadata": {}, + "source": [ + "Read at face value, the Laplace approximation puts $\\theta_{\\rm true}$ almost twenty standard deviations away, and $\\theta^\\ast$ more than four. The first is no surprise: $\\theta_{\\rm true}$ differs from $\\theta^\\ast$ by the part of $\\delta$ that the line can imitate, and nothing in any likelihood can recover it. The second says the mode does not hit $\\theta^\\ast$ either. Under the Kennedy–O'Hagan prior, the mode lands wherever the fitted kernel happens to split the line-like part of the residual between $\\delta$ and $\\theta$; change the kernel family or its priors, and it moves. The estimate has no well-defined target. (We will also see, when we sample, that this error bar is far too small.)\n", + "\n", + "## Plumlee's orthogonal discrepancy\n", + "\n", + "[Plumlee (2017)](https://doi.org/10.1080/01621459.2017.1329149) removes the ambiguity by forbidding the discrepancy from containing anything the model can do. The prior on $\\delta$ is conditioned on\n", + "\n", + "$$\\int_0^{10} \\delta(x)\\, \\frac{\\partial\\eta}{\\partial\\theta}(x)\\, dx = 0,$$\n", + "\n", + "which for our line means $\\delta$ is $L^2$-orthogonal to both $1$ and $x$. Conditioning a GP on a linear functional is again a GP, with the kernel\n", + "\n", + "$$k^\\perp(x, x') = k(x, x') - g(x)^\\top M^{-1} g(x'), \\qquad g(x) = \\int k(x, u)\\, h(u)\\, du, \\qquad M = \\iint h(u)\\, k(u, u')\\, h(u')^\\top du\\, du'.$$\n", + "\n", + "Under this prior, $\\theta$ *is* $\\theta^\\ast$ by construction.\n", + "\n", + "It is still a kernel of $x$, so it is still an sklearn kernel and still a `T.kernel` term, and it still travels to a prediction grid. We orthogonalise each component separately, so the decomposition into components stays exact. The integrals use [Gauss–Legendre](https://en.wikipedia.org/wiki/Gauss%E2%80%93Legendre_quadrature) quadrature with a node count that follows the component's spectrum, since the integrand oscillates at $\\mu$ under an envelope of width $s$; a fixed rule accurate for the drift would alias the oscillation." + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "id": "f938c755", + "metadata": {}, + "outputs": [], + "source": [ + "@lru_cache(maxsize=64)\n", + "def gauss_legendre(n_nodes, lo, hi):\n", + " t, wt = np.polynomial.legendre.leggauss(n_nodes)\n", + " return 0.5 * (hi - lo) * (t + 1) + lo, 0.5 * (hi - lo) * wt\n", + "\n", + "\n", + "class OrthogonalSpectralComponent(SpectralComponent):\n", + " \"\"\"A spectral component conditioned on int delta h dx = 0 over [lo, hi].\"\"\"\n", + "\n", + " def __init__(\n", + " self, mu=0.5, s=0.3, mu_bounds=(1e-3, 1e2), s_bounds=(1e-3, 1e2), lo=0.0, hi=L\n", + " ):\n", + " super().__init__(mu, s, mu_bounds, s_bounds)\n", + " self.lo, self.hi = lo, hi\n", + "\n", + " def __call__(self, X, Y=None, eval_gradient=False):\n", + " if eval_gradient:\n", + " raise NotImplementedError(\"rxmc never asks a kernel for gradients\")\n", + " X = np.atleast_2d(X)\n", + " Y = X if Y is None else np.atleast_2d(Y)\n", + " width = self.hi - self.lo\n", + " n_nodes = int(\n", + " np.clip(np.ceil(4 * width * (self.mu + 2 * self.s)) + 16, 32, 400)\n", + " )\n", + " u, wu = gauss_legendre(n_nodes, self.lo, self.hi)\n", + " U, Hw = u[:, None], wu[:, None] * h(u)\n", + " gX, gY = self._k(X, U) @ Hw, self._k(Y, U) @ Hw\n", + " M = Hw.T @ self._k(U, U) @ Hw\n", + " return self._k(X, Y) - gX @ np.linalg.solve(M, gY.T)\n", + "\n", + " def diag(self, X):\n", + " return np.diag(self(X))\n", + "\n", + " def is_stationary(self):\n", + " return False\n", + "\n", + "\n", + "plumlee_terms = components(OrthogonalSpectralComponent)\n", + "plumlee = rx.Problem([rx.Constraint([comp], terms=plumlee_terms)])" + ] + }, + { + "cell_type": "markdown", + "id": "642367dd", + "metadata": {}, + "source": [ + "Everything else is unchanged: the same line, the same data, the same priors and bounds, the same starts." + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "id": "88c4d89a", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "component 0: amplitude 0.137 period 9.57 envelope 5.06\n", + "component 1: amplitude 0.351 period 0.71 envelope 3.37\n", + "component 2: amplitude 0.024 period 0.42 envelope 4.48\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Plumlee: theta_hat = [2.204 0.447] +/- [0.01 0.002]\n", + " Mahalanobis distance to theta_true: 33.97\n", + " Mahalanobis distance to theta* : 2.20\n", + "CPU times: user 5min 6s, sys: 57.8 ms, total: 5min 6s\n", + "Wall time: 5min 9s\n" + ] + } + ], + "source": [ + "%%time\n", + "plumlee_map = fit_map(plumlee)\n", + "describe(plumlee, plumlee_terms, plumlee_map)\n", + "plumlee_theta, plumlee_cov = report(plumlee, plumlee_map, \"Plumlee\")" + ] + }, + { + "cell_type": "markdown", + "id": "9770d806", + "metadata": {}, + "source": [ + "The orthogonal fit lands on $\\theta^\\ast$, about two standard deviations away, with a smaller error bar than the Kennedy–O'Hagan fit. It is no closer to $\\theta_{\\rm true}$, and could not be: orthogonality changes what $\\theta$ *means*, it does not add information. What it buys is an estimator with a definite target, and error bars that describe the distance to that target.\n", + "\n", + "The components rearrange too: one slow component, one fast one carrying almost all of the oscillation, and a third with less than a tenth of its amplitude. Below we compare the *total* discrepancy each fit reconstructs." + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "id": "a06f5bd8", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "plumlee_bands = component_bands(plumlee, plumlee_terms, plumlee_map)\n", + "delta_perp = delta_true[n:] - h(x_fine) @ (theta_star - theta_true)\n", + "fig, ax = plt.subplots()\n", + "ax.plot(\n", + " x_fine,\n", + " delta_true[n:],\n", + " \":\",\n", + " color=\"0.4\",\n", + " label=r\"true $\\delta$ about $\\theta_{\\rm true}$\",\n", + ")\n", + "ax.plot(\n", + " x_fine,\n", + " delta_perp,\n", + " color=\"k\",\n", + " lw=1,\n", + " label=r\"true $\\delta^\\perp$ about $\\theta^\\ast$\",\n", + ")\n", + "for i, (label, bands) in enumerate(\n", + " [(\"Kennedy-O'Hagan\", koh_bands), (\"Plumlee\", plumlee_bands)]\n", + "):\n", + " ax.plot(\n", + " x_fine,\n", + " sum(b[1] for b in bands),\n", + " color=plotstyle.COLOURS[i],\n", + " label=rf\"{label}: $\\sum_q \\hat\\delta_q$\",\n", + " )\n", + "ax.set(xlabel=\"$x$\", ylabel=r\"$\\delta(x)$\")\n", + "ax.legend(fontsize=8)\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "f3da09cd", + "metadata": {}, + "source": [ + "Both fits reconstruct the *total* discrepancy about their own $\\hat\\theta$ well, and both follow $\\delta^\\perp$ rather than the $\\delta$ we drew: the data never knew which line we started from.\n", + "\n", + "## Checking the Laplace approximation with a chain\n", + "\n", + "Recipe 33 says to compare a Laplace covariance against a chain before reporting it. We run an [emcee](https://emcee.readthedocs.io/) chain from a small ball around each mode and print its integrated autocorrelation time $\\tau$, the number of steps between effectively independent samples." + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "id": "b54ad503", + "metadata": {}, + "outputs": [], + "source": [ + "def chain(problem, t, n_steps=1500, n_walkers=24, seed=0):\n", + " lo, hi = problem.bounds[:, 0], problem.bounds[:, 1]\n", + " r = np.random.default_rng(seed)\n", + " start = np.clip(\n", + " t + 1e-3 * r.standard_normal((n_walkers, t.size)), lo + 1e-9, hi - 1e-9\n", + " )\n", + " sampler = emcee.EnsembleSampler(n_walkers, problem.ndim, problem.log_posterior)\n", + " sampler.random_state = np.random.RandomState(seed).get_state()\n", + " sampler.run_mcmc(start, n_steps, progress=False)\n", + " tau = sampler.get_autocorr_time(tol=0).max() # a lower bound from a short chain\n", + " print(\n", + " f\"acceptance {sampler.acceptance_fraction.mean():.2f}, \"\n", + " f\"longest tau {tau:.0f} steps: {n_steps / tau:.0f} tau in the chain\"\n", + " )\n", + " return sampler.get_chain(discard=500, thin=5, flat=True)" + ] + }, + { + "cell_type": "code", + "execution_count": 18, + "id": "792a9461", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "acceptance 0.28, longest tau 141 steps: 11 tau in the chain\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "acceptance 0.29, longest tau 168 steps: 9 tau in the chain\n", + "CPU times: user 6min 53s, sys: 119 ms, total: 6min 53s\n", + "Wall time: 6min 58s\n" + ] + } + ], + "source": [ + "%%time\n", + "koh_chain = chain(koh, koh_map)\n", + "plumlee_chain = chain(plumlee, plumlee_map)" + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "id": "684b9047", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Kennedy-O'Hagan chain [2.084 0.462] +/- [0.1731 0.0252]\n", + " Laplace [2.147 0.452] +/- [0.0226 0.0045]\n", + "Plumlee chain [2.202 0.447] +/- [0.0102 0.0018]\n", + " Laplace [2.204 0.447] +/- [0.01 0.0017]\n" + ] + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "cols = koh.columns(line.params)\n", + "for label, ch, th, C in [\n", + " (\"Kennedy-O'Hagan\", koh_chain, koh_theta, koh_cov),\n", + " (\"Plumlee\", plumlee_chain, plumlee_theta, plumlee_cov),\n", + "]:\n", + " s = ch[:, cols]\n", + " print(f\"{label:16s} chain {s.mean(0).round(3)} +/- {s.std(0).round(4)}\")\n", + " print(f\"{'':16s} Laplace {th.round(3)} +/- {np.sqrt(np.diag(C)).round(4)}\")\n", + "\n", + "\n", + "def ellipse(ax, mean, cov, k, **kwargs):\n", + " \"\"\"The k-sigma ellipse of a 2-D Gaussian.\"\"\"\n", + " vals, vecs = np.linalg.eigh(cov)\n", + " angle = np.degrees(np.arctan2(vecs[1, 1], vecs[0, 1]))\n", + " width, height = 2 * k * np.sqrt(vals[1]), 2 * k * np.sqrt(vals[0])\n", + " ax.add_patch(Ellipse(mean, width, height, angle=angle, fill=False, **kwargs))\n", + "\n", + "\n", + "fig, (wide, zoom) = plt.subplots(1, 2, figsize=(9.0, 4.0))\n", + "runs = [\n", + " (\"Kennedy-O'Hagan\", koh_chain, koh_theta, koh_cov, \"--\"),\n", + " (\"Plumlee\", plumlee_chain, plumlee_theta, plumlee_cov, \"-\"),\n", + "]\n", + "for ax in (wide, zoom):\n", + " for i, (label, ch, th, C, ls) in enumerate(runs):\n", + " ax.plot(*ch[:, cols].T, \".\", ms=1.5, alpha=0.3, color=plotstyle.COLOURS[i])\n", + " ellipse(ax, th, C, 2, color=\"k\", ls=ls, lw=1.5, zorder=3)\n", + " ax.plot(*theta_true, \"*\", color=\"k\", ms=12)\n", + " ax.plot(*theta_star, \"X\", color=\"k\", ms=8)\n", + " ax.set(xlabel=r\"$\\theta_0$\", ylabel=r\"$\\theta_1$\")\n", + "for i, (label, _, _, _, ls) in enumerate(runs): # legend proxies\n", + " wide.plot([], [], \"o\", ms=4, color=plotstyle.COLOURS[i], label=f\"{label}: chain\")\n", + " wide.plot([], [], ls, color=\"k\", lw=1.5, label=f\"{label}: Laplace, 2 sd\")\n", + "wide.plot([], [], \"*\", color=\"k\", ms=10, label=r\"$\\theta_{\\rm true}$\")\n", + "wide.plot([], [], \"X\", color=\"k\", ms=7, label=r\"$\\theta^\\ast$\")\n", + "wide.legend(fontsize=7)\n", + "wide.set(title=\"both chains\")\n", + "zoom.set(\n", + " xlim=theta_star[0] + np.array([-0.12, 0.08]),\n", + " ylim=theta_star[1] + np.array([-0.012, 0.02]),\n", + " title=r\"zoomed on $\\theta^\\ast$\",\n", + ")\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "id": "0cc8b631", + "metadata": {}, + "source": [ + "The two models behave completely differently.\n", + "\n", + "Both chains are short: about ten autocorrelation times each, and an autocorrelation time estimated from a short chain is itself an underestimate. The hyperparameters mix slowly in both. What that costs depends on how much $\\theta$ cares about them.\n", + "\n", + "- **Plumlee**: the chain and the Laplace ellipse agree, and both sit on $\\theta^\\ast$. With the discrepancy forbidden from imitating the line, $\\theta$ barely moves as the hyperparameters wander, so it is sharply identified and the Laplace approximation is good enough.\n", + "- **Kennedy–O'Hagan**: the chain is several times wider than the Laplace ellipse, stretched along a ridge of anticorrelated intercept and slope, and it reaches $\\theta_{\\rm true}$. Being short, it *understates* the width: a run four times longer with more walkers (not executed here) has an autocorrelation time near 400 steps and a standard deviation of 0.28 in $\\theta_0$, more than ten times the Laplace value. The data fix $\\theta$ only up to what the slow component can absorb, and the posterior spreads along that trade-off, with its extent set by the prior on $\\delta$ rather than by the data.\n", + "\n", + "So the Kennedy–O'Hagan posterior covers $\\theta_{\\rm true}$, but not because it learned anything about it: it is wide because $\\theta$ is not identified. A MAP-and-Laplace summary of the same posterior was wrong twice, in location and in width.\n", + "\n", + "\n", + "## What to take from this\n", + "\n", + "- **A spectral mixture detects scales, not mechanisms.** Two scales an order of magnitude apart, each inside the band the design resolves, came out as two distinct spectral modes. A scale near the window length is detected but its spectral parameters are not meaningful, and a scale shorter than the point spacing would be indistinguishable from noise without replicates or a known experimental covariance.\n", + "- **Surplus components do not reliably switch off.** Here the third component split the oscillation with another rather than vanishing. A half-Cauchy amplitude prior allows a component to vanish but does not force it to. Count distinct spectral modes, not active components, or use a prior that favours few components.\n", + "- **The calibrated $\\theta$ is the projection $\\theta^\\ast$ at best.** Under Kennedy–O'Hagan, $\\theta$ has no defined target: its posterior is wide, slow to sample, and shaped by the prior on $\\delta$, and a Laplace approximation at the mode hides all three. Plumlee's orthogonality gives $\\theta$ a target, and its error bars then describe the distance to it. Recovering more of $\\theta_{\\rm true}$ needs information about $\\delta$ in the span of $\\partial\\eta/\\partial\\theta$, which the data cannot provide: a physics prior on the form of the defect, or several observables that share $\\theta$ but not $\\delta$.\n", + "- **The measure in the orthogonality condition is a modelling choice.** $\\theta^\\ast$ depends on the domain and its weighting: uniform in angle, uniform in $\\cos\\theta$ and the experimental design density give three different \"physical\" parameters.\n", + "- **A nonlinear model makes $k^\\perp$ depend on $\\theta$**, since $h = \\partial\\eta/\\partial\\theta$ does. The kernel must then be rebuilt at each $\\theta$, which is natural when the model is differentiable and $h$ is its Jacobian." + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.13.17" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/src/rxmc/covariance.py b/src/rxmc/covariance.py index 36739cb..5aa9dde 100644 --- a/src/rxmc/covariance.py +++ b/src/rxmc/covariance.py @@ -29,6 +29,19 @@ ``B`` must be positive definite. A block covered only by modes is singular in ``B`` even when ``Sigma`` is not; when every term is constant this is caught at construction and reported with the block's label. + +A covariance that is not finite at a ``theta`` (a term returning inf or NaN, +or values whose squares, sums or low-rank products overflow) is zero density +there: :meth:`distance` returns a non-finite ``d2`` or ``logdet`` and the +compiled constraint reads it as ``log_likelihood == -inf``. The likelihood +path never scans terms for this: the factorisations run unchecked and a +non-finite input always yields a non-finite ``logdet``. A constant term is +checked once, at construction, and a non-finite contribution raises +``ValueError`` naming it. :meth:`matrix` checks too and raises, because a +predictive has no "zero density" to return. One gap remains: constant terms +whose contributions are each finite but whose sum overflows, in a block a +parametric term also touches, are zero density everywhere rather than an +error at construction. """ from __future__ import annotations @@ -49,9 +62,16 @@ class _SingularCovariance(ValueError): """ -def chol_logdet(Sigma): - """Lower Cholesky factor and log-determinant of a positive-definite matrix.""" - L = sla.cholesky(np.asarray(Sigma, dtype=float), lower=True) +def chol_logdet(Sigma, check_finite=True): + """Lower Cholesky factor and log-determinant of a positive-definite matrix. + + ``check_finite=False`` skips scipy's scan: a non-finite ``Sigma`` then gives + a non-finite factor and log-determinant (or ``LinAlgError``) instead of + ``ValueError``. + """ + L = sla.cholesky( + np.asarray(Sigma, dtype=float), lower=True, check_finite=check_finite + ) return L, 2.0 * float(np.sum(np.log(np.diag(L)))) @@ -156,12 +176,14 @@ def __init__(self, entries, x, y, offsets, active, *, meta=None, labels=None): self.constant_entries = [e for e in self.entries if e.term.is_constant] self.parametric_entries = [e for e in self.entries if not e.term.is_constant] self.is_constant = not self.parametric_entries - self._D0, self._U0, self._M0, self._X0 = self._pieces( - self.constant_entries, None, None - ) + with np.errstate(over="ignore", invalid="ignore"): + self._D0, self._U0, self._M0, self._X0 = self._pieces( + self.constant_entries, None, None, check=True + ) self._cache = None if self.is_constant: - self._factor(self._D0, self._U0, self._M0, self._X0) + with np.errstate(over="ignore", invalid="ignore"): + self._factor(self._D0, self._U0, self._M0, self._X0) elif not self.dense: # modes never enter B, so a block no parametric diag or matrix term # touches has a constant B: a singular one fails now, by label, rather @@ -170,8 +192,12 @@ def __init__(self, entries, x, y, offsets, active, *, meta=None, labels=None): # -- assembly ------------------------------------------------------------- - def _pieces(self, entries, ym, theta): - """``(D, U_columns, M_blocks, cross_blocks)`` from a set of entries.""" + def _pieces(self, entries, ym, theta, check=False): + """``(D, U_columns, M_blocks, cross_blocks)`` from a set of entries. + + ``check`` raises ``ValueError`` naming the first term whose contribution + is not finite on its active rows; the likelihood path leaves it off. + """ n = self.n_active D = np.zeros(n) U = [] @@ -190,14 +216,20 @@ def _pieces(self, entries, ym, theta): labels=e.labels, ) pos, keep = e.pos[e.keep], e.keep + if t.kind == "matrix": + part = v[np.ix_(keep, keep)] + else: + part = v[keep] ** 2 if t.kind == "diag" else v[keep] + if check and not np.all(np.isfinite(part)): + self._nonfinite(t, theta is None) if t.kind == "diag": - D[pos] += v[keep] ** 2 + D[pos] += part elif t.kind == "mode": col = np.zeros(n) - col[pos] = v[keep] + col[pos] = part U.append(col) else: - sub = v[np.ix_(keep, keep)] + sub = part if e.block is None: X.append((pos, sub)) else: @@ -211,11 +243,11 @@ def _pieces(self, entries, ym, theta): M[b][np.ix_(local, local)] += sub return D, U, M, X - def _assemble(self, ym, theta): + def _assemble(self, ym, theta, check=False): """Constant pieces plus the parametric ones at ``theta``.""" if self.is_constant: return self._D0, self._U0, self._M0, self._X0 - D, U, M, X = self._pieces(self.parametric_entries, ym, theta) + D, U, M, X = self._pieces(self.parametric_entries, ym, theta, check) D = D + self._D0 U = list(self._U0) + U M = [ @@ -237,12 +269,17 @@ def _dense_matrix(self, D, U, M, X): # -- factorisation --------------------------------------------------------- def _factor(self, D, U, M, X): - """Factor the covariance; cached when constant.""" + """Factor the covariance; cached when constant. + + Unchecked: a non-finite input gives a non-finite ``logdet`` (or a + ``LinAlgError``, depending on where LAPACK meets it), both of which the + compiled constraint reads as zero density. + """ if self.is_constant and self._cache is not None: return self._cache try: if self.dense: - L, logdet = chol_logdet(self._dense_matrix(D, U, M, X)) + L, logdet = chol_logdet(self._dense_matrix(D, U, M, X), False) factors = ("dense", L, logdet) else: blocks = [] @@ -255,11 +292,14 @@ def _factor(self, D, U, M, X): B = np.diag(D[pos]) if M[b] is not None: B = B + M[b] - L, ld = chol_logdet(B) + L, ld = chol_logdet(B, False) logdet += ld W = ( sla.solve_triangular( - L, np.column_stack([c[pos] for c in U]), lower=True + L, + np.column_stack([c[pos] for c in U]), + lower=True, + check_finite=False, ) if U else np.zeros((pos.size, 0)) @@ -268,7 +308,7 @@ def _factor(self, D, U, M, X): Ws.append(W) if U: W = np.vstack(Ws) if Ws else np.zeros((0, len(U))) - Ls, lds = chol_logdet(np.eye(len(U)) + W.T @ W) + Ls, lds = chol_logdet(np.eye(len(U)) + W.T @ W, False) logdet += lds else: Ls = None @@ -277,9 +317,33 @@ def _factor(self, D, U, M, X): error = ValueError if self.is_constant else _SingularCovariance raise error(self._singular_message(D)) from err if self.is_constant: + if not np.isfinite(logdet): + raise ValueError( + "the constraint's constant covariance is not finite: its " + "terms are finite one by one but overflow when combined " + "(e.g. a mode of magnitude ~1e200)" + ) self._cache = factors return factors + def _nonfinite(self, term, constant: bool): + """Raise ``ValueError``: ``term`` contributed inf or NaN on its active rows.""" + msg = f"{term!r} contributes a non-finite value on its active rows" + if constant: + raise ValueError( + msg + "; a constant term is evaluated once, so this is an error " + "in its declaration (an array with NaN, values whose squares " + "overflow, or a fn that divides by zero on the data)" + ) + raise ValueError(msg + " at this theta") + + def has_nan(self, ym, theta) -> bool: + """Whether the assembled covariance at ``theta`` contains NaN (diagnostic).""" + with np.errstate(over="ignore", invalid="ignore"): + D, U, M, X = self._assemble(ym, np.asarray(theta, dtype=float)) + arrays = [D, *U, *(m for m in M if m is not None), *(x for _, x in X)] + return any(np.isnan(a).any() for a in arrays) + def _constant_blocks(self) -> list[int]: """The blocks whose ``B`` no parametric diag or matrix entry touches.""" varying = set() @@ -328,40 +392,55 @@ def distance(self, ym, theta=()): r"""``(d2, logdet)`` of the residual ``y - ym`` on the active rows. ``ym`` is the full-length stacked prediction; ``theta`` the flat - parameter vector the entries' gathers index into. + parameter vector the entries' gathers index into. Either value is + non-finite where the covariance (or the residual) overflows: zero + density, read as such by the compiled constraint. """ theta = np.asarray(theta, dtype=float) r = (self.y - np.asarray(ym, dtype=float))[self.active] - kind, *rest = self._factor(*self._assemble(ym, theta)) - if kind == "dense": - L, logdet = rest - z = sla.solve_triangular(L, r, lower=True) - return float(z @ z), float(logdet) - blocks, Ls, logdet = rest - d2 = 0.0 - w = None - for blk in blocks: - if blk is None: - continue - pos, L, W = blk - z = sla.solve_triangular(L, r[pos], lower=True) - d2 += float(z @ z) - w = W.T @ z if w is None else w + W.T @ z - if Ls is not None and w is not None: - s = sla.solve_triangular(Ls, w, lower=True) - d2 -= float(s @ s) - return d2, float(logdet) - - def matrix(self, ym, theta=(), entries=None): - """The dense covariance on the active rows, for display and tests. + with np.errstate(over="ignore", invalid="ignore"): + kind, *rest = self._factor(*self._assemble(ym, theta)) + if kind == "dense": + L, logdet = rest + z = sla.solve_triangular(L, r, lower=True, check_finite=False) + return float(z @ z), float(logdet) + blocks, Ls, logdet = rest + d2 = 0.0 + w = None + for blk in blocks: + if blk is None: + continue + pos, L, W = blk + z = sla.solve_triangular(L, r[pos], lower=True, check_finite=False) + d2 += float(z @ z) + w = W.T @ z if w is None else w + W.T @ z + if Ls is not None and w is not None: + s = sla.solve_triangular(Ls, w, lower=True, check_finite=False) + d2 -= float(s @ s) + return d2, float(logdet) + + def matrix(self, ym, theta=(), entries=None, check=True): + """The dense covariance on the active rows, for display and predictives. ``entries`` restricts the sum to a subset of :attr:`entries` (the objects themselves, compared by identity): the covariance of *part* of the error model, as a predictive draw of the model plus its discrepancy but not the experimental errors needs. Default: every entry. A subset is assembled directly, without the constant-piece cache. + + Raises ``ValueError`` if the matrix is not finite, naming the term when + one term alone is responsible; ``check=False`` returns it as is. """ theta = np.asarray(theta, dtype=float) - if entries is None: - return self._dense_matrix(*self._assemble(ym, theta)) - return self._dense_matrix(*self._pieces(entries, ym, theta)) + with np.errstate(over="ignore", invalid="ignore"): + if entries is None: + pieces = self._assemble(ym, theta, check=check) + else: + pieces = self._pieces(entries, ym, theta, check=check) + Sigma = self._dense_matrix(*pieces) + if check and not np.all(np.isfinite(Sigma)): + raise ValueError( + "the covariance is not finite at this theta: every term is " + "finite on its own, but their sum or a low-rank product overflows" + ) + return Sigma diff --git a/src/rxmc/diagnostics.py b/src/rxmc/diagnostics.py index 85be867..2d3f03e 100644 --- a/src/rxmc/diagnostics.py +++ b/src/rxmc/diagnostics.py @@ -107,6 +107,24 @@ def _masks(constraint) -> list[np.ndarray]: return list(constraint.source.masks) +def _row_error(problem, i, theta, reason) -> ValueError: + """The error a predictive raises at a sample row with no density.""" + at = ", ".join(f"{n}={v:.6g}" for n, v in zip(problem.names, theta)) + return ValueError( + f"sample row {i} ({at}): {reason}. This point has zero likelihood, so a " + "posterior chain cannot contain it: check burn-in and column order. In " + "a prior predictive, bound the prior where the error model is undefined." + ) + + +def _require_finite(problem, i, theta, *arrays) -> None: + """Raise :func:`_row_error` unless every array is finite.""" + if not all(np.all(np.isfinite(a)) for a in arrays): + raise _row_error( + problem, i, theta, "the prediction or covariance is not finite" + ) + + class _Conditional: """``p(y_H | y_F, theta)`` for one held-out constraint given its fit.""" @@ -151,9 +169,16 @@ def __init__(self, held: Problem, given: Problem, constraint: int): self.yF = self.full.y[f.active] def __call__(self, theta): - """``(mean, cov)`` of the held-out rows given the fitted data.""" + """``(mean, cov)`` of the held-out rows given the fitted data. + + All NaN where the prediction or covariance is not finite (zero density): + the scorer returns ``-inf`` there and a predictive raises. + """ ym = self.full.ym(theta) - S = self.full.covariance.matrix(ym, theta) + S = self.full.covariance.matrix(ym, theta, check=False) + if not (np.all(np.isfinite(ym[self.full.active])) and np.all(np.isfinite(S))): + n = len(self.iH) + return np.full(n, np.nan), np.full((n, n), np.nan) SHH = S[np.ix_(self.iH, self.iH)] SHF = S[np.ix_(self.iH, self.iF)] SFF = S[np.ix_(self.iF, self.iF)] @@ -264,6 +289,7 @@ def predictive_draws( out = np.empty((n, N)) for i in range(n): out[i] = cond(samples[i])[0] if cond else c.ym(samples[i])[c.active] + _require_finite(problem, i, samples[i], out[i]) return out if return_draws else np.percentile(out, levels, axis=0) out = np.empty((n * n_rep, N)) @@ -273,7 +299,11 @@ def predictive_draws( mu, Sigma = cond(theta) else: mu = c.ym(theta)[c.active] - Sigma = c.matrix(theta, terms=terms, statistical=statistical) + try: + Sigma = c.matrix(theta, terms=terms, statistical=statistical) + except ValueError as err: # names the term + raise _row_error(problem, i, theta, str(err)) from err + _require_finite(problem, i, theta, mu, Sigma) L = _psd_factor(Sigma) z = rng.standard_normal((n_rep, N)) z *= c.likelihood.predictive_scale(rng, n_rep, *theta[c.like_gather])[:, None] @@ -371,6 +401,9 @@ def heldout_log_predictive( if c.weight == 0.0: continue mean, cov = cond(theta) + if not np.all(np.isfinite(mean)): # no density at this theta + total = -np.inf + break total += c.weight * _gaussian_logpdf(c.y[c.active] - mean, cov) out[k] = total return out diff --git a/src/rxmc/predictive.py b/src/rxmc/predictive.py index 17c4f88..0e9c556 100644 --- a/src/rxmc/predictive.py +++ b/src/rxmc/predictive.py @@ -77,7 +77,7 @@ import numpy as np import scipy.linalg as sla -from .diagnostics import _psd_factor, _rows +from .diagnostics import _psd_factor, _require_finite, _row_error, _rows from .problem import Problem, _per_point from .terms import KernelTerm, TermContext, as_2d @@ -331,6 +331,7 @@ def _draw_on_grid( out = np.empty((n, n_pred)) for i, theta in enumerate(samples): out[i] = back(space(predictor(*theta[cols_pred]))) + _require_finite(problem, i, theta, out[i]) return out if return_draws else predictive_band(out, levels) out = np.empty((n * n_rep, n_pred)) @@ -341,9 +342,13 @@ def _draw_on_grid( C += _grid_piece(t, x_pred, mu, meta_p, tuple(theta[g])) f_mean = 0.0 if condition is not None: - f_mean, v = condition(theta, mu) + try: + f_mean, v = condition(theta, mu) + except ValueError as err: # the training covariance names the term + raise _row_error(problem, i, theta, str(err)) from err C -= v.T @ v C[np.diag_indices(n_pred)] += float(noise_std) ** 2 + _require_finite(problem, i, theta, mu, f_mean, C) z = rng.standard_normal((n_rep, n_pred)) if joint: z = z @ _psd_factor(0.5 * (C + C.T)).T diff --git a/src/rxmc/problem.py b/src/rxmc/problem.py index 208a523..6d557a7 100644 --- a/src/rxmc/problem.py +++ b/src/rxmc/problem.py @@ -402,6 +402,7 @@ class CompiledConstraint: def __init__(self, constraint: Constraint, index: ParameterIndex): self.source = constraint + self._warned_nan = False comps = constraint.comparisons self.comparisons = comps self.labels = [c.data.label or f"comparison {i}" for i, c in enumerate(comps)] @@ -473,13 +474,38 @@ def predict_physical(self, theta) -> list[np.ndarray]: return [c.predictor(*theta[g]) for _, g, c in self.predictors] def _stats(self, theta): + """``(d2, logdet)``, or ``None`` where the error model has no density.""" ym = self.ym(theta) if not np.all(np.isfinite(ym[self.active])): return None try: - return self.covariance.distance(ym, theta) - except _SingularCovariance: - return None # a parametric covariance singular at this theta: no density + d2, logdet = self.covariance.distance(ym, theta) + except _SingularCovariance: # singular at this theta + self._warn_if_nan(ym, theta) + return None + if not (np.isfinite(d2) and np.isfinite(logdet)): # overflow or NaN + self._warn_if_nan(ym, theta) + return None + return d2, logdet + + def _warn_if_nan(self, ym, theta): + """Warn once per constraint when a rejected covariance contains NaN. + + Runs only after a rejection, so accepted points pay nothing. NaN can be + a term bug or ``inf * 0`` from an overflowing amplitude; either way the + point is zero density, and :meth:`matrix` names the term. + """ + if self._warned_nan or not self.covariance.has_nan(ym, theta): + return + self._warned_nan = True + warnings.warn( + f"constraint {self.labels}: the covariance is NaN at theta=" + f"{np.array2string(np.asarray(theta), precision=4)} (a term bug, or " + "inf*0 from an overflowing amplitude); such points get zero density. " + "constraint.matrix(theta) names the term. Warned once per constraint.", + RuntimeWarning, + stacklevel=4, + ) def log_likelihood(self, theta) -> float: s = self._stats(theta) diff --git a/src/rxmc/terms.py b/src/rxmc/terms.py index bb31130..c65b508 100644 --- a/src/rxmc/terms.py +++ b/src/rxmc/terms.py @@ -70,6 +70,7 @@ "proportional_error", "systematic", "kernel", + "kernel_param_names", ] KINDS = ("diag", "mode", "matrix") @@ -216,7 +217,7 @@ def __post_init__(self): if self.kind == "matrix": if a.ndim != 2 or a.shape[0] != a.shape[1]: raise ValueError(f"matrix term expects a square block, got {a.shape}") - if not np.allclose(a, a.T): + if not np.allclose(a, a.T, equal_nan=True): raise ValueError("matrix term must be symmetric") elif a.ndim != 1: raise ValueError(f"{self.kind} term expects a vector, got shape {a.shape}") @@ -291,7 +292,7 @@ def __repr__(self): return f"Term(kind={self.kind!r}, params=({names}), on={self.on!r})" -@dataclass(eq=False, frozen=True) +@dataclass(eq=False, frozen=True, repr=False) class KernelTerm(Term): """A :func:`kernel` term that also carries what GP conditioning needs. @@ -562,8 +563,88 @@ def _kernel_params(kernel, prefix) -> list: return params -def _n_free_elements(kernel) -> int: - return sum(hp.n_elements for hp in kernel.hyperparameters if not hp.fixed) +def kernel_param_names(kernel) -> list[str]: + """The names of a kernel's free hyperparameters, in ``kernel.theta`` order. + + These are the keys :func:`kernel` accepts in a ``params=`` mapping. The + order is sklearn's, which sorts by the ``hyperparameter_*`` attribute + name, so it is alphabetical and unrelated to the order of ``__init__`` + arguments; a composite kernel prefixes its operands' names (``k1__``, + ``k2__``). A hyperparameter with ``n_elements > 1`` (an anisotropic length + scale) appears once and takes a sequence of that many parameters. + """ + return [hp.name for hp in kernel.hyperparameters if not hp.fixed] + + +def _ordered(v, what) -> list: + """A Parameter, list or tuple as a list; anything unordered is refused. + + A set would reorder the elements of an anisotropic hyperparameter by hash. + """ + if isinstance(v, Parameter): + return [v] + if not isinstance(v, (list, tuple)): + raise TypeError( + f"{what} must be a Parameter, list or tuple (ordered), got " + f"{type(v).__name__}" + ) + return list(v) + + +def _resolve_kernel_params(kernel, params) -> list: + """``params=`` as a list of :class:`Parameter` in ``kernel.theta`` order. + + A mapping is keyed by :func:`kernel_param_names`. A sequence is accepted + only when the kernel has a single free hyperparameter: with two or more, + the meaning of a position depends on sklearn's alphabetical ordering, and a + misread order swaps priors without any error, so it is refused. + """ + free = [hp for hp in kernel.hyperparameters if not hp.fixed] + names = [hp.name for hp in free] + nk = kernel.n_dims + if isinstance(params, Mapping): + fixed = [hp.name for hp in kernel.hyperparameters if hp.fixed] + missing = [n for n in names if n not in params] + unknown = [k for k in params if k not in names and k not in fixed] + given_fixed = [k for k in params if k in fixed] + if missing or unknown or given_fixed: + raise ValueError( + f"params= keys must be the kernel's free hyperparameters {names} " + f"(kernel.theta order); missing {missing}, unknown {unknown}" + + (f", fixed (not sampled) {given_fixed}" if given_fixed else "") + ) + out = [] + for hp in free: + v = params[hp.name] + if hp.n_elements == 1: + if not isinstance(v, Parameter): + raise TypeError( + f"params[{hp.name!r}] must be a Parameter, got {v!r}" + ) + out.append(v) + else: + v = _ordered(v, f"params[{hp.name!r}]") + if len(v) != hp.n_elements: + raise ValueError( + f"params[{hp.name!r}] needs {hp.n_elements} Parameters " + f"(one per element), got {len(v)}" + ) + out.extend(v) + return out + out = _ordered(params, "params=") + if len(free) > 1: + raise TypeError( + f"the kernel has {len(free)} free hyperparameters {names}; pass " + "params= as a mapping {name: Parameter}. A positional list is " + "refused because sklearn orders hyperparameters alphabetically by " + "attribute name, so a misread order would silently swap priors" + ) + if len(out) != nk: + raise ValueError( + f"kernel has {nk} free hyperparameter element(s) but params= " + f"has {len(out)}" + ) + return out def kernel( @@ -609,23 +690,23 @@ def kernel( Name prefix of the derived kernel parameters. Two kernel terms with derived names and the same prefix fail to compile on the duplicate name: pass distinct prefixes, or ``params=`` to share one kernel. - params : sequence of Parameter, optional - The hyperparameter objects themselves, one per free element in - ``kernel.theta`` order. Pass the same objects to several ``kernel`` - calls to share hyperparameters between comparisons. + params : mapping or sequence of Parameter, optional + The hyperparameter objects themselves, keyed by hyperparameter name + (:func:`kernel_param_names`): ``{"length_scale": ell}``. A + multi-element hyperparameter takes a sequence of that many objects. + A plain sequence is accepted only for a kernel with a single free + hyperparameter; with two or more it is refused, because the position + of each follows sklearn's alphabetical ordering and a misread order + would silently swap priors. Pass the same objects to several + ``kernel`` calls to share hyperparameters between comparisons. on : optional See :class:`Term`. """ - nk = _n_free_elements(kernel) + nk = kernel.n_dims # free hyperparameter elements (len(kernel.theta)) if params is None: kparams = _kernel_params(kernel, prefix) else: - kparams = list(params) - if len(kparams) != nk: - raise ValueError( - f"kernel has {nk} free hyperparameter element(s) but params= " - f"has {len(kparams)}" - ) + kparams = _resolve_kernel_params(kernel, params) amplitude_params = tuple(amplitude_params) fixed_K = nk == 0 and not as_transform(coords).params diff --git a/test/test_covariance.py b/test/test_covariance.py index 886cbf7..3528306 100644 --- a/test/test_covariance.py +++ b/test/test_covariance.py @@ -2,6 +2,7 @@ import numpy as np import pytest +from scipy import stats from sklearn.gaussian_process.kernels import RBF, Matern from helpers import ( @@ -13,7 +14,15 @@ ) from rxmc import Parameter from rxmc.covariance import StructuredCovariance, chol_logdet -from rxmc.terms import Term, kernel, noise, normalization, offset, statistical +from rxmc.terms import ( + Term, + constant_amplitude, + kernel, + noise, + normalization, + offset, + statistical, +) def build(terms, x, y, offsets, active=None, rows=None, labels=None): @@ -324,3 +333,128 @@ def test_masked_rows_stay_in_the_segment_view(self): # so the segments describe the unmasked support seen = self.capture(np.arange(9), active=np.r_[0:2, 3:9]) assert seen["segments"] == (slice(0, 3), slice(3, 6), slice(6, 9)) + + +class TestNonFinite: + """A non-finite covariance is zero density at a theta, an error at compile.""" + + def setup_method(self): + from rxmc import Comparison, Dataset, Model + + x = np.linspace(0.0, 2.0, 6) + self.m = Parameter("m", prior=stats.norm(0, 5)) + model = Model(lambda x, m: m * x, [self.m]) + self.comps = [ + Comparison( + Dataset(x, 2.0 * x + 0.1 * k, 0.1 * np.ones(6), label=f"d{k}"), model + ) + for k in range(2) + ] + + def problem(self, terms, comps=None, masks=None): + from rxmc import Constraint, Problem + + c = Constraint(comps or self.comps[:1], terms=terms) + return Problem([c if masks is None else c.masked(masks)]) + + def assert_no_density(self, p, theta): + assert p.log_likelihood(theta) == -np.inf + assert p.log_posterior(theta) == -np.inf + assert p.chi2(theta) == np.inf + + def amplitude_kernel(self): + la = Parameter("log_A", prior=stats.norm(0, 1)) + return kernel( + RBF(0.5, "fixed"), amplitude=constant_amplitude, amplitude_params=(la,) + ) + + # -- the likelihood: -inf, and no numpy warnings --------------------------- + + @pytest.mark.filterwarnings("error") + def test_overflowing_kernel_amplitude_structured_path(self): + p = self.problem([self.amplitude_kernel()]) + assert not p.constraints[0].covariance.dense + assert np.isfinite(p.log_likelihood(np.array([2.0, 0.0]))) + self.assert_no_density(p, np.array([2.0, 400.0])) # exp(400)**2 overflows + + @pytest.mark.filterwarnings("error") + def test_overflowing_kernel_amplitude_dense_path(self): + p = self.problem([self.amplitude_kernel()], comps=self.comps) # dense + assert p.constraints[0].covariance.dense + self.assert_no_density(p, np.array([2.0, 400.0])) + + @pytest.mark.filterwarnings("error") + def test_finite_values_whose_square_overflows(self): + # exp(400) is finite; the diagonal adds its square, which is not + p = self.problem([noise(Parameter("log_eps", prior=stats.norm(0, 1)))]) + self.assert_no_density(p, np.array([2.0, 400.0])) + + @pytest.mark.filterwarnings("error") + def test_overflowing_residual_on_the_woodbury_path(self): + # every term finite; z@z and s@s overflow and d2 = inf - inf is NaN + from rxmc import Comparison, Dataset, Model + + m = Parameter("m", prior=stats.norm(0, 200)) + x = np.linspace(0.0, 2.0, 6) + comp = Comparison( + Dataset(x, 2.0 * x, 0.1 * np.ones(6)), + Model(lambda x, m: np.exp(m) * x, [m]), + ) + p = self.problem([offset(magnitude=10.0)], comps=[comp]) + assert np.isfinite(p.log_likelihood(np.array([0.5]))) + self.assert_no_density(p, np.array([370.0])) + + def test_parametric_nan_is_zero_density_and_warns_once(self): + import warnings + + s = Parameter("s", prior=stats.norm(0, 1)) + t = Term(lambda c, s: np.sqrt(s) * np.ones(len(c)), (s,), kind="diag") + p = self.problem([t]) + assert np.isfinite(p.log_likelihood(np.array([2.0, 0.04]))) + with warnings.catch_warnings(record=True) as caught: + warnings.simplefilter("always") + for _ in range(10): + self.assert_no_density(p, np.array([2.0, -1.0])) + # one named warning; numpy's per-call "invalid value in sqrt" is silenced + assert len(caught) == 1 + assert issubclass(caught[0].category, RuntimeWarning) + assert "NaN" in str(caught[0].message) and "d0" in str(caught[0].message) + + # -- construction: a constant term is checked once and named --------------- + + def test_constant_term_with_nan_fails_at_compile_and_names_it(self): + bad = np.full(6, 0.1) + bad[2] = np.nan + with pytest.raises(ValueError, match="kind='diag'.*non-finite.*declaration"): + self.problem([Term(bad, kind="diag")]) + + @pytest.mark.parametrize("kind", ["mode", "matrix"]) + def test_constant_mode_or_matrix_with_nan_names_it(self, kind): + bad = np.full(6, 0.1) if kind == "mode" else 0.01 * np.eye(6) + bad[(2,) if kind == "mode" else (2, 4)] = np.nan + if kind == "matrix": + bad[4, 2] = np.nan + with pytest.raises(ValueError, match=f"kind='{kind}'.*non-finite"): + self.problem([Term(bad, kind=kind)]) + + def test_nan_on_a_masked_row_compiles(self): + bad = np.full(6, 0.1) + bad[2] = np.nan + mask = np.ones(6, bool) + mask[2] = False + p = self.problem([Term(bad, kind="diag")], masks=[mask]) + assert np.isfinite(p.log_likelihood(np.array([2.0]))) + + def test_constant_square_overflow_raises_beside_a_parametric_term(self): + terms = [ + Term(np.full(6, 1e200), kind="diag"), + noise(Parameter("log_eps", prior=stats.norm(0, 1))), + ] + with pytest.raises(ValueError, match="non-finite.*declaration"): + self.problem(terms) + + @pytest.mark.filterwarnings("error") + def test_constant_mode_whose_product_overflows_raises(self): + # the mode column is finite; W^T W is not + with pytest.raises(ValueError, match="overflow when combined"): + self.problem([Term(np.full(6, 1e200), kind="mode")]) diff --git a/test/test_notebooks_index.py b/test/test_notebooks_index.py index af18dda..241b93a 100644 --- a/test/test_notebooks_index.py +++ b/test/test_notebooks_index.py @@ -16,7 +16,7 @@ EXAMPLES = ROOT / "examples" RECIPES = ROOT / "docs" / "recipes.md" -# design document section 9, item 8: the nine notebooks and their recipes +# design document section 6: the notebooks and the recipes each cites NOTEBOOKS = { "linear_calibration": {1, 2, 17, 40}, "error_models": {2, 4, 19}, @@ -28,6 +28,7 @@ "local_optical_model_calibration": {10, 14, 15, 16, 17, 18, 19, 21, 40}, "alpha_ca_error_model_comparison": {7, 10, 13, 17, 18, 19, 40}, "hierarchical_calibration": {22, 24, 30, 35, 38}, + "spectral_mixture_discrepancy": {7, 33, 40}, } RECIPE_LINE = re.compile(r"^Recipes?:\s*(.+)$", re.M) diff --git a/test/test_predictive.py b/test/test_predictive.py index 36c04da..7f93374 100644 --- a/test/test_predictive.py +++ b/test/test_predictive.py @@ -592,3 +592,78 @@ def test_point_by_point_terms_have_no_value_at_a_new_x(self): grid_draws(p, pred, X_GRID, theta, terms=[normalization(magnitude=0.1)]) with pytest.raises(ValueError, match=r"\(n, 3\)"): grid_draws(p, pred, X_GRID, np.zeros((4, 2))) + + +class TestInvalidRows: + """A sample row with no density: predictives raise, scoring gives -inf.""" + + def setup_method(self): + from rxmc import Constraint, Problem + from rxmc.terms import constant_amplitude + + x = np.linspace(0.0, 2.0, 8) + self.m = Parameter("m", prior=stats.norm(0, 5)) + self.la = Parameter("log_A", prior=stats.norm(0, 1)) + self.model = Model(lambda x, m: m * x, [self.m]) + self.data = Dataset(x, 2.0 * x, 0.1 * np.ones(8), label="d") + self.comp = Comparison(self.data, self.model) + self.gp = kernel( + RBF(0.5, "fixed"), + amplitude=constant_amplitude, + amplitude_params=(self.la,), + on=self.comp, + ) + self.c = Constraint([self.comp], terms=[self.gp]) + self.p = Problem([self.c]) + self.good = np.array([2.0, np.log(0.1)]) + self.bad = np.array([2.0, 400.0]) # the amplitude overflows + + def test_matrix_raises_a_plain_value_error_naming_the_term(self): + with pytest.raises(ValueError, match="kind=.matrix.*log_A.*non-finite") as err: + self.p.constraints[0].matrix(self.bad) + assert type(err.value) is ValueError + assert len(str(err.value)) < 300 # KernelTerm uses Term's short repr + + @pytest.mark.parametrize("model_only", [False, True]) + def test_predictive_draws_name_the_row_and_values(self, model_only): + samples = np.vstack([self.good, self.bad]) + if model_only: # a model that is not finite at the second row + samples[1] = [np.inf, 0.0] + with pytest.raises(ValueError, match=r"sample row 1 \(m=.*log_A="): + predictive_draws(self.p, samples, model_only=model_only, rng=0) + assert np.all(np.isfinite(predictive_draws(self.p, samples[:1], rng=0))) + + def test_grid_draws_raise_on_the_row(self): + x_new = np.linspace(0.0, 3.0, 5) + pred = self.model.bind(x_new, self.data.meta) + with pytest.raises(ValueError, match="sample row 1"): + grid_draws( + self.p, pred, x_new, np.vstack([self.good, self.bad]), terms=[self.gp] + ) + + def test_heldout_scoring_is_minus_inf_with_or_without_given(self): + from rxmc import Problem + from rxmc.diagnostics import heldout_log_predictive + + mask = np.arange(8) < 5 + fit = self.c.masked([mask]) + held = Problem([fit.complement()]) + samples = np.vstack([self.good, self.bad]) + for given in (None, Problem([fit])): + lp = heldout_log_predictive(held, samples, given=given) + assert np.isfinite(lp[0]) and lp[1] == -np.inf + + def test_matrix_raises_when_only_a_product_overflows(self): + from rxmc import Constraint, Problem + from rxmc.terms import offset + + log_w = Parameter("log_w", prior=stats.norm(0, 1)) + p = Problem([Constraint([self.comp], terms=[offset(log_w)])]) + # the mode exp(360) ~ 1e156 is finite; its outer product is not + with pytest.raises(ValueError, match="low-rank product overflows"): + p.constraints[0].matrix(np.array([2.0, 360.0])) + + def test_predictive_draws_raise_on_a_bad_prediction_with_a_finite_covariance(self): + samples = np.vstack([self.good, [np.inf, np.log(0.1)]]) + with pytest.raises(ValueError, match="sample row 1.*not finite"): + predictive_draws(self.p, samples, rng=0) diff --git a/test/test_terms.py b/test/test_terms.py index 1d8894f..78c7788 100644 --- a/test/test_terms.py +++ b/test/test_terms.py @@ -2,7 +2,13 @@ import numpy as np import pytest -from sklearn.gaussian_process.kernels import RBF, ConstantKernel, Matern, WhiteKernel +from sklearn.gaussian_process.kernels import ( + RBF, + ConstantKernel, + ExpSineSquared, + Matern, + WhiteKernel, +) from helpers import STUDY_LEGEND, assemble_dense, study_form from rxmc import Parameter @@ -15,6 +21,7 @@ exp_growth, exp_growth_amplitude, kernel, + kernel_param_names, noise, normalization, offset, @@ -357,6 +364,90 @@ def test_shared_hyperparameters_via_params(self): with pytest.raises(ValueError, match="free hyperparameter"): kernel(RBF(1.0), params=[ell, Parameter("extra")]) + def test_param_names_are_sklearn_theta_order(self): + # sklearn sorts by attribute name, not by __init__ order + k = ConstantKernel(1.0) * ExpSineSquared(1.0, 2.0) + assert kernel_param_names(k) == [ + "k1__constant_value", + "k2__length_scale", + "k2__periodicity", + ] + assert kernel_param_names(RBF(1.0, length_scale_bounds="fixed")) == [] + + def test_mapping_params_follow_theta_order_whatever_the_key_order(self): + k = ExpSineSquared(1.0, 2.0) + ell, per = Parameter("ell"), Parameter("per") + t1 = kernel(k, params={"length_scale": ell, "periodicity": per}) + t2 = kernel(k, params={"periodicity": per, "length_scale": ell}) + assert t1.params == t2.params == (ell, per) + x = np.linspace(0.0, 3.0, 6) + vals = (np.log(0.7), np.log(1.3)) + S1 = t1.value(x, np.zeros(6), np.zeros(6), *vals) + S2 = t2.value(x, np.zeros(6), np.zeros(6), *vals) + ref = ExpSineSquared(0.7, 1.3)(x[:, None]) + np.testing.assert_allclose(S1, S2) + np.testing.assert_allclose(S1, ref + 1e-10 * np.eye(6) * np.mean(np.diag(ref))) + + def test_positional_params_refused_for_several_hyperparameters(self): + # the hazard: ExpSineSquared's theta is (length_scale, periodicity); + # a reader expecting (periodicity, length_scale) would swap the priors + with pytest.raises(TypeError, match="mapping"): + kernel(ExpSineSquared(1.0, 2.0), params=[Parameter("a"), Parameter("b")]) + + def test_mapping_keys_are_validated(self): + k = ExpSineSquared(1.0, 2.0) + with pytest.raises(ValueError, match=r"missing \['periodicity'\]"): + kernel(k, params={"length_scale": Parameter("a")}) + with pytest.raises(ValueError, match=r"unknown \['period'\]"): + kernel( + k, + params={ + "length_scale": Parameter("a"), + "periodicity": Parameter("b"), + "period": Parameter("c"), + }, + ) + with pytest.raises(TypeError, match="must be a Parameter"): + kernel(k, params={"length_scale": 1.0, "periodicity": Parameter("b")}) + + def test_mapping_with_a_multi_element_hyperparameter(self): + k = ConstantKernel(1.0) * RBF(length_scale=[1.0, 1.0]) + A, l1, l2 = Parameter("A"), Parameter("l1"), Parameter("l2") + t = kernel(k, params={"k1__constant_value": A, "k2__length_scale": [l1, l2]}) + assert t.params == (A, l1, l2) + with pytest.raises(ValueError, match="needs 2 Parameters"): + kernel(k, params={"k1__constant_value": A, "k2__length_scale": [l1]}) + + def test_unordered_containers_are_refused(self): + ps = [Parameter(f"l{i}") for i in range(4)] + with pytest.raises(TypeError, match="ordered.*set"): + kernel(RBF([1.0] * 4), params=set(ps)) + with pytest.raises(TypeError, match="ordered.*set"): + kernel(RBF([1.0] * 4), params={"length_scale": set(ps)}) + with pytest.raises(TypeError, match="ordered.*dict"): + kernel(RBF([1.0] * 2), params={"length_scale": {ps[0]: 1, ps[1]: 2}}) + with pytest.raises(TypeError, match="ordered.*generator"): + kernel(RBF(1.0), params=(p for p in ps[:1])) + with pytest.raises(TypeError, match="ordered.*float"): + kernel(RBF(1.0), params=1.0) + + def test_a_fixed_hyperparameter_key_is_named_as_fixed(self): + k = ConstantKernel(1.0, "fixed") * RBF(1.0) + with pytest.raises(ValueError, match=r"fixed \(not sampled\) \['k1__"): + kernel( + k, + params={ + "k1__constant_value": Parameter("a"), + "k2__length_scale": Parameter("b"), + }, + ) + + def test_positional_is_fine_for_one_hyperparameter_even_anisotropic(self): + l1, l2 = Parameter("l1"), Parameter("l2") + t = kernel(RBF(length_scale=[1.0, 1.0]), params=[l1, l2]) + assert t.params == (l1, l2) + assert kernel(RBF(1.0), params=l1).params == (l1,) + def test_constant_amplitude_reproduces_constant_kernel(self): x = np.linspace(0.0, 2.0, 5) A, la = 0.7, Parameter("log A")