# Copyright (C) 2017-2022 Cleanlab Inc.
# This file is part of cleanlab.
#
# cleanlab is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as published
# by the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# cleanlab is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with cleanlab. If not, see <https://www.gnu.org/licenses/>.
"""
Contains mathematical functions relating the latent terms,
``P(given_label)``, ``P(given_label | true_label)``, ``P(true_label | given_label)``, ``P(true_label)``, etc. together.
For every function here, if the inputs are exact, the output is guaranteed to be exact.
Every function herein is the computational equivalent of a mathematical equation having a closed, exact form.
If the inputs are inexact, the error will of course propagate.
Throughout `K` denotes the number of classes in the classification task.
"""
import warnings
import numpy as np
from typing import Tuple
from cleanlab.internal.util import value_counts, clip_values, clip_noise_rates, TINY_VALUE
[docs]def compute_ps_py_inv_noise_matrix(
labels, noise_matrix
) -> Tuple[np.ndarray, np.ndarray, np.ndarray]:
"""Compute ``ps := P(labels=k), py := P(true_labels=k)``, and the inverse noise matrix.
Parameters
----------
labels : np.ndarray
A discrete vector of noisy labels, i.e. some labels may be erroneous.
*Format requirements*: for dataset with `K` classes, labels must be in ``{0,1,...,K-1}``.
noise_matrix : np.ndarray
A conditional probability matrix (of shape ``(K, K)``) of the form ``P(label=k_s|true_label=k_y)`` containing
the fraction of examples in every class, labeled as every other class.
Assumes columns of noise_matrix sum to 1."""
ps = value_counts(labels) / float(len(labels)) # p(labels=k)
py, inverse_noise_matrix = compute_py_inv_noise_matrix(ps, noise_matrix)
return ps, py, inverse_noise_matrix
[docs]def compute_py_inv_noise_matrix(ps, noise_matrix) -> Tuple[np.ndarray, np.ndarray]:
"""Compute py := P(true_label=k), and the inverse noise matrix.
Parameters
----------
ps : np.ndarray
Array of shape ``(K, )`` or ``(1, K)``.
The fraction (prior probability) of each observed, NOISY class ``P(labels = k)``.
noise_matrix : np.ndarray
A conditional probability matrix (of shape ``(K, K)``) of the form ``P(label=k_s|true_label=k_y)`` containing
the fraction of examples in every class, labeled as every other class.
Assumes columns of noise_matrix sum to 1."""
# 'py' is p(true_labels=k) = noise_matrix^(-1) * p(labels=k)
# because in *vector computation*: P(label=k|true_label=k) * p(true_label=k) = P(label=k)
# The pseudo-inverse is used when noise_matrix is not invertible.
py = np.linalg.inv(noise_matrix).dot(ps)
# No class should have probability 0, so we use .000001
# Make sure valid probabilities that sum to 1.0
py = clip_values(py, low=1e-6, high=1.0, new_sum=1.0)
# All the work is done in this function (below)
return py, compute_inv_noise_matrix(py=py, noise_matrix=noise_matrix, ps=ps)
[docs]def compute_inv_noise_matrix(py, noise_matrix, *, ps=None) -> np.ndarray:
"""Compute the inverse noise matrix if py := P(true_label=k) is given.
Parameters
----------
py : np.ndarray (shape (K, 1))
The fraction (prior probability) of each TRUE class label, P(true_label = k)
noise_matrix : np.ndarray
A conditional probability matrix (of shape ``(K, K)``) of the form ``P(label=k_s|true_label=k_y)`` containing
the fraction of examples in every class, labeled as every other class.
Assumes columns of noise_matrix sum to 1.
ps : np.ndarray
Array of shape ``(K, 1)`` containing the fraction (prior probability) of each NOISY given label, ``P(labels = k)``.
`ps` is easily computable from py and should only be provided if it has already been precomputed, to increase code efficiency.
Examples
--------
For loop based implementation:
.. code:: python
# Number of classes
K = len(py)
# 'ps' is p(labels=k) = noise_matrix * p(true_labels=k)
# because in *vector computation*: P(label=k|true_label=k) * p(true_label=k) = P(label=k)
if ps is None:
ps = noise_matrix.dot(py)
# Estimate the (K, K) inverse noise matrix P(true_label = k_y | label = k_s)
inverse_noise_matrix = np.empty(shape=(K,K))
# k_s is the class value k of noisy label `label == k`
for k_s in range(K):
# k_y is the (guessed) class value k of true label y
for k_y in range(K):
# P(true_label|label) = P(label|y) * P(true_label) / P(labels)
inverse_noise_matrix[k_y][k_s] = noise_matrix[k_s][k_y] * \
py[k_y] / ps[k_s]
"""
joint = noise_matrix * py
ps = joint.sum(axis=1) if ps is None else ps
inverse_noise_matrix = joint.T / np.clip(ps, a_min=TINY_VALUE, a_max=None)
# Clip inverse noise rates P(true_label=k_s|true_label=k_y) into proper range [0,1)
return clip_noise_rates(inverse_noise_matrix)
[docs]def compute_noise_matrix_from_inverse(ps, inverse_noise_matrix, *, py=None) -> np.ndarray:
"""Compute the noise matrix ``P(label=k_s|true_label=k_y)``.
Parameters
----------
py : np.ndarray
Array of shape ``(K, 1)`` containing the fraction (prior probability) of each TRUE class label, ``P(true_label = k)``.
inverse_noise_matrix : np.ndarray
A conditional probability matrix (of shape ``(K, K)``) of the form P(true_label=k_y|label=k_s) representing
the estimated fraction observed examples in each class k_s, that are
mislabeled examples from every other class k_y. If None, the
inverse_noise_matrix will be computed from pred_probs and labels.
Assumes columns of inverse_noise_matrix sum to 1.
ps : np.ndarray
Array of shape ``(K, 1)`` containing the fraction (prior probability) of each observed NOISY label, P(labels = k).
`ps` is easily computable from `py` and should only be provided if it has already been precomputed, to increase code efficiency.
Returns
-------
noise_matrix : np.ndarray
Array of shape ``(K, K)``, where `K` = number of classes, whose columns sum to 1.
A conditional probability matrix of the form ``P(label=k_s|true_label=k_y)`` containing
the fraction of examples in every class, labeled as every other class.
Examples
--------
For loop based implementation:
.. code:: python
# Number of classes labels
K = len(ps)
# 'py' is p(true_label=k) = inverse_noise_matrix * p(true_label=k)
# because in *vector computation*: P(true_label=k|label=k) * p(label=k) = P(true_label=k)
if py is None:
py = inverse_noise_matrix.dot(ps)
# Estimate the (K, K) noise matrix P(labels = k_s | true_labels = k_y)
noise_matrix = np.empty(shape=(K,K))
# k_s is the class value k of noisy label `labels == k`
for k_s in range(K):
# k_y is the (guessed) class value k of true label y
for k_y in range(K):
# P(labels|y) = P(true_label|labels) * P(labels) / P(true_label)
noise_matrix[k_s][k_y] = inverse_noise_matrix[k_y][k_s] * \
ps[k_s] / py[k_y]
"""
joint = (inverse_noise_matrix * ps).T
py = joint.sum(axis=0) if py is None else py
noise_matrix = joint / np.clip(py, a_min=TINY_VALUE, a_max=None)
# Clip inverse noise rates P(true_label=k_y|true_label=k_s) into proper range [0,1)
return clip_noise_rates(noise_matrix)
[docs]def compute_py(
ps, noise_matrix, inverse_noise_matrix, *, py_method="cnt", true_labels_class_counts=None
) -> np.ndarray:
"""Compute ``py := P(true_labels=k)`` from ``ps := P(labels=k)``, `noise_matrix`, and
`inverse_noise_matrix`.
This method is ** ROBUST ** when ``py_method = 'cnt'``
It may work well even when the noise matrices are estimated
poorly by using the diagonals of the matrices
instead of all the probabilities in the entire matrix.
Parameters
----------
ps : np.ndarray
Array of shape ``(K, )`` or ``(1, K)`` containing the fraction (prior probability) of each observed, noisy label, P(labels = k)
noise_matrix : np.ndarray
A conditional probability matrix ( of shape ``(K, K)``) of the form ``P(label=k_s|true_label=k_y)`` containing
the fraction of examples in every class, labeled as every other class.
Assumes columns of noise_matrix sum to 1.
inverse_noise_matrix : np.ndarray of shape (K, K), K = number of classes
A conditional probability matrix ( of shape ``(K, K)``) of the form ``P(true_label=k_y|label=k_s)`` representing
the estimated fraction observed examples in each class `k_s`, that are
mislabeled examples from every other class `k_y`. If ``None``, the
inverse_noise_matrix will be computed from `pred_probs` and `labels`.
Assumes columns of `inverse_noise_matrix` sum to 1.
py_method : str (Options: ["cnt", "eqn", "marginal", "marginal_ps"])
How to compute the latent prior ``p(true_label=k)``. Default is "cnt" as it often
works well even when the noise matrices are estimated poorly by using
the matrix diagonals instead of all the probabilities.
true_labels_class_counts : np.ndarray
Array of shape ``(K, )`` or ``(1, K)`` containing the marginal counts of the confident joint
(like ``cj.sum(axis = 0)``).
Returns
-------
py : np.ndarray
Array of shape ``(K, )`` or ``(1, K)``.
The fraction (prior probability) of each TRUE class label, ``P(true_label = k)``."""
if len(np.shape(ps)) > 2 or (len(np.shape(ps)) == 2 and np.shape(ps)[0] != 1):
w = "Input parameter np.ndarray ps has shape " + str(np.shape(ps))
w += ", but shape should be (K, ) or (1, K)"
warnings.warn(w)
if py_method == "marginal" and true_labels_class_counts is None:
msg = (
'py_method == "marginal" requires true_labels_class_counts, '
"but true_labels_class_counts is None. "
)
msg += " Provide parameter true_labels_class_counts."
raise ValueError(msg)
if py_method == "cnt":
# Computing py this way avoids dividing by zero noise rates.
# More robust bc error est_p(true_label|labels) / est_p(labels|y) ~ p(true_label|labels) / p(labels|y)
py = (
inverse_noise_matrix.diagonal()
/ np.clip(noise_matrix.diagonal(), a_min=TINY_VALUE, a_max=None)
* ps
)
# Equivalently: py = (true_labels_class_counts / labels_class_counts) * ps
elif py_method == "eqn":
py = np.linalg.inv(noise_matrix).dot(ps)
elif py_method == "marginal":
py = true_labels_class_counts / np.clip(
float(sum(true_labels_class_counts)), a_min=TINY_VALUE, a_max=None
)
elif py_method == "marginal_ps":
py = np.dot(inverse_noise_matrix, ps)
else:
err = "py_method {}".format(py_method)
err += " should be in [cnt, eqn, marginal, marginal_ps]"
raise ValueError(err)
# Clip py (0,1), s.t. no class should have prob 0, hence 1e-5
py = clip_values(py, low=1e-5, high=1.0, new_sum=1.0)
return py
[docs]def compute_pyx(pred_probs, noise_matrix, inverse_noise_matrix):
"""Compute ``pyx := P(true_label=k|x)`` from ``pred_probs := P(label=k|x)``, `noise_matrix` and
`inverse_noise_matrix`.
This method is ROBUST - meaning it works well even when the
noise matrices are estimated poorly by only using the diagonals of the
matrices which tend to be easy to estimate correctly.
Parameters
----------
pred_probs : np.ndarray
``P(label=k|x)`` is a ``(N x K)`` matrix with K model-predicted probabilities.
Each row of this matrix corresponds to an example `x` and contains the model-predicted
probabilities that `x` belongs to each possible class.
The columns must be ordered such that these probabilities correspond to class 0,1,2,...
`pred_probs` should have been computed using 3 (or higher) fold cross-validation.
noise_matrix : np.ndarray
A conditional probability matrix (of shape ``(K, K)``) of the form ``P(label=k_s|true_label=k_y)`` containing
the fraction of examples in every class, labeled as every other class.
Assumes columns of `noise_matrix` sum to 1.
inverse_noise_matrix : np.ndarray
A conditional probability matrix (of shape ``(K, K)``) of the form ``P(true_label=k_y|label=k_s)`` representing
the estimated fraction observed examples in each class `k_s`, that are
mislabeled examples from every other class `k_y`. If None, the
inverse_noise_matrix will be computed from `pred_probs` and `labels`.
Assumes columns of `inverse_noise_matrix` sum to 1.
Returns
-------
pyx : np.ndarray
``P(true_label=k|x)`` is a ``(N, K)`` matrix of model-predicted probabilities.
Each row of this matrix corresponds to an example `x` and contains the model-predicted
probabilities that `x` belongs to each possible class.
The columns must be ordered such that these probabilities correspond to class 0,1,2,...
`pred_probs` should have been computed using 3 (or higher) fold cross-validation."""
if len(np.shape(pred_probs)) != 2:
raise ValueError(
"Input parameter np.ndarray 'pred_probs' has shape "
+ str(np.shape(pred_probs))
+ ", but shape should be (N, K)"
)
pyx = (
pred_probs
* inverse_noise_matrix.diagonal()
/ np.clip(noise_matrix.diagonal(), a_min=TINY_VALUE, a_max=None)
)
# Make sure valid probabilities that sum to 1.0
return np.apply_along_axis(
func1d=clip_values, axis=1, arr=pyx, **{"low": 0.0, "high": 1.0, "new_sum": 1.0}
)