cleanlab is a machine learning python package for learning with noisy labels and finding label errors in datasets. cleanlab CLEANs LABels. It is powered by the theory of confident learning, published in this paper | blog.

News! (Dec 2020) cleanlab supports NeurIPS research. (Northcutt et al., NeurIPS, 2020), and finds label errors across 10 common benchark datasets (ImageNet, CIFAR-10, CIFAR-100, Caltech-256, Quickdraw, MNIST, Amazon Reviews, IMDB, 20 News Groups, AudioSet). Along with the paper, the authors launched labelerrors.com where you can view the label errors in these datasets.
News! (Dec 2020) cleanlab supports PU learning.
News! (Jan 2020) cleanlab achieves state-of-the-art on CIFAR-10 for learning with noisy labels. Code to reproduce is here: examples/cifar10. This is a great place for newcomers to see how to use cleanlab on real datasets. Data needed is available in the confidentlearning-reproduce repo, cleanlab v0.1.0 reproduces results in the CL paper.
News! (Feb 2020) cleanlab now natively supports Mac, Linux, and Windows.
News! (Feb 2020) cleanlab now supports Co-Teaching (Han et al., 2018).

cleanlab documentation is available in this blog post.

Past release notes and future features planned is available here.

So fresh, so `cleanlab`

cleanlab finds and cleans label errors in any dataset using state-of-the-art algorithms to find label errors, characterize noise, and learn in spite of it. cleanlab is fast: its built on optimized algorithms and parallelized across CPU threads automatically. cleanlab is powered by provable guarantees of exact noise estimation and label error finding in realistic cases when model output probabilities are erroneous. cleanlab supports multi-label, multiclass, sparse matrices, etc. By default, cleanlab requires no hyper-parameters.

cleanlab implements the family of theory and algorithms called confident learning with provable guarantees of exact noise estimation and label error finding (even when model output probabilities are noisy/imperfect).

How does confident learning work? See: TUTORIAL: confident learning with just numpy and for-loops.

cleanlab supports most weak supervision tasks: multi-label, multiclass, sparse matrices, etc.

cleanlab is:

backed-by-theory - Provable perfect label error finding in realistic conditions.
fast - Non-iterative, parallelized algorithms (e.g. < 1 second to find label errors in ImageNet)
general - Works with any ML or deep learning framework: PyTorch, Tensorflow, MxNet, Caffe2, scikit-learn, etc.
unique - The only package for weak supervion with any dataset / classifier.

Find label errors with PyTorch, Tensorflow, MXNet, etc. in 1 line of code.

# Compute psx (n x m matrix of predicted probabilities) on your own, with any classifier.
# Here is an example that shows in detail how to compute psx on CIFAR-10:
#    https://github.com/cgnorthcutt/cleanlab/tree/master/examples/cifar10
# Be sure you compute probs in a holdout/out-of-sample manner (e.g. cross-validation)
# Now getting label errors is trivial with cleanlab... its one line of code.
# Label errors are ordered by likelihood of being an error. First index is most likely error.
from cleanlab.pruning import get_noise_indices

ordered_label_errors = get_noise_indices(
    s=numpy_array_of_noisy_labels,
    psx=numpy_array_of_predicted_probabilities,
    sorted_index_method='normalized_margin', # Orders label errors
 )

Pre-computed out-of-sample predicted probabilities for CIFAR-10 train set are available here: [[LINK]].

Learning with noisy labels in 3 lines of code!

from cleanlab.classification import LearningWithNoisyLabels
from sklearn.linear_model import LogisticRegression

# Wrap around any classifier. Yup, you can use sklearn/pyTorch/Tensorflow/FastText/etc.
lnl = LearningWithNoisyLabels(clf=LogisticRegression()) 
lnl.fit(X=X_train_data, s=train_noisy_labels) 
# Estimate the predictions you would have gotten by training with *no* label errors.
predicted_test_labels = lnl.predict(X_test)

Check out these examples and tests (includes how to use pyTorch, FastText, etc.).

Installation

Python 2.7, 3.4, 3.5, 3.6, and 3.7 are supported. Linux, macOS, and Windows are supported.

Stable release:

$ pip install cleanlab

Developer (unstable) release:

$ pip install git+https://github.com/cgnorthcutt/cleanlab.git

To install the codebase (enabling you to make modifications):

$ conda update pip # if you use conda
$ git clone https://github.com/cgnorthcutt/cleanlab.git
$ cd cleanlab
$ pip install -e .

If you use this package in your work, please cite the confident learning paper:

@misc{northcutt2019confidentlearning,
   title={Confident Learning: Estimating Uncertainty in Dataset Labels},
   author={Curtis G. Northcutt and Lu Jiang and Isaac L. Chuang},
   year={2019},
   eprint={1911.00068},
   archivePrefix={arXiv},
   primaryClass={stat.ML}
}

If used for binary classification, cleanlab also implements this paper:

@inproceedings{northcutt2017rankpruning,
 author={Northcutt, Curtis G. and Wu, Tailin and Chuang, Isaac L.},
 title={Learning with Confident Examples: Rank Pruning for Robust Classification with Noisy Labels},
 booktitle = {Proceedings of the Thirty-Third Conference on Uncertainty in Artificial Intelligence},
 series = {UAI'17},
 year = {2017},
 location = {Sydney, Australia},
 numpages = {10},
 url = {http://auai.org/uai2017/proceedings/papers/35.pdf},
 publisher = {AUAI Press},
}

Reproducing Results in confident learning paper

See cleanlab/examples/cifar10 and cleanlab/examples/imagenet. You'll need to git clone confidentlearning-reproduce which contains the data and files needed to reproduce the CIFAR-10 results.

`cleanlab`: State of the Art Learning with Noisy Labels in CIFAR

A [step-by-step guide] to reproduce these results is available [here]. This guide is also helpful as a tutorial to use cleanlab on any large-scale dataset.

Comparison of confident learning (CL) and cleanlab versus seven recent methods for learning with noisy labels in CIFAR-10. Highlighted cells show CL robustness to sparsity. The five CL methods estimate label errors, remove them, then train on the cleaned data using Co-Teaching.

Observe how cleanlab (CL methods) are robust to large sparsity in label noise whereas prior art tends to reduce in performance for increased sparsity, as shown by the red highlighted regions. This is important because real-world label noise is often sparse, e.g. a tiger is likely to be mislabeled as a lion, but not as most other classes like airplane, bathtub, and microwave.

`cleanlab`: Find Label Errors in ImageNet

Use cleanlab to identify ~100,000 label errors in the 2012 ImageNet training dataset.

Top label issues in the 2012 ILSVRC ImageNet train set identified using cleanlab. Label Errors are boxed in red. Ontological issues in green. Multi-label images in blue.

`cleanlab`: Find Label Errors in MNIST

Use cleanlab to identify ~50 label errors in the MNIST dataset.

Label errors of the original MNIST train dataset identified algorithmically using cleanlab. Depicts the 24 least confident labels, ordered left-right, top-down by increasing self-confidence (probability of belonging to the given label), denoted conf in teal. The label with the largest predicted probability is in green. Overt errors are in red.

`cleanlab` Generality: View performance across 4 distributions and 9 classifiers.

Use cleanlab to learn with noisy labels regardless of dataset distribution or classifier.

Each sub-figure in the figure above depicts the decision boundary learned using cleanlab.classification.LearningWithNoisyLabels in the presence of extreme (~35%) label errors. Label errors are circled in green. Label noise is class-conditional (not simply uniformly random). Columns are organized by the classifier used, except the left-most column which depicts the ground-truth dataset distribution. Rows are organized by dataset used.

The code to reproduce this figure is available here.

Each figure depicts accuracy scores on a test set as decimal values:

LEFT (in black): The classifier test accuracy trained with perfect labels (no label errors).
MIDDLE (in blue): The classifier test accuracy trained with noisy labels using cleanlab.
RIGHT (in white): The baseline classifier test accuracy trained with noisy labels.

As an example, this is the noise matrix (noisy channel) P(s | y) characterizing the label noise for the first dataset row in the figure. s represents the observed noisy labels and y represents the latent, true labels. The trace of this matrix is 2.6. A trace of 4 implies no label noise. A cell in this matrix is read like, "A random 38% of '3' labels were flipped to '2' labels."

p(s\|y)	y=0	y=1	y=2	y=3
s=0	0.55	0.01	0.07	0.06
s=1	0.22	0.87	0.24	0.02
s=2	0.12	0.04	0.64	0.38
s=3	0.11	0.08	0.05	0.54

Get started with easy, quick examples.

New to cleanlab? Start with:

These examples show how easy it is to characterize label noise in datasets, learn with noisy labels, identify label errors, estimate latent priors and noisy channels, and more.

Use `cleanlab` with any model (Tensorflow, caffe2, PyTorch, etc.)

All of the features of the cleanlab package work with any model. Yes, any model. Feel free to use PyTorch, Tensorflow, caffe2, scikit-learn, mxnet, etc. If you use a scikit-learn classifier, all cleanlab methods will work out-of-the-box. It’s also easy to use your favorite model from a non-scikit-learn package, just wrap your model into a Python class that inherits the sklearn.base.BaseEstimator:

from sklearn.base import BaseEstimator
class YourFavoriteModel(BaseEstimator): # Inherits sklearn base classifier
    def __init__(self, ):
        pass
    def fit(self, X, y, sample_weight=None):
        pass
    def predict(self, X):
        pass
    def predict_proba(self, X):
        pass
    def score(self, X, y, sample_weight=None):
        pass

# Now you can use your model with `cleanlab`. Here's one example:
from cleanlab.classification import LearningWithNoisyLabels
lnl = LearningWithNoisyLabels(clf=YourFavoriteModel())
lnl.fit(train_data, train_labels_with_errors)

Want to see a working example? Here’s a compliant PyTorch MNIST CNN class

As you can see here, technically you don’t actually need to inherit from sklearn.base.BaseEstimator, as you can just create a class that defines .fit(), .predict(), and .predict_proba(), but inheriting makes downstream scikit-learn applications like hyper-parameter optimization work seamlessly. For example, the LearningWithNoisyLabels() model is fully compliant.

Note, some libraries exists to do this for you. For pyTorch, check out the skorch Python library which will wrap your pytorch model into a scikit-learn compliant model.

Documentation by Example

`cleanlab` Core Package Components

cleanlab/classification.py - The LearningWithNoisyLabels() class for learning with noisy labels.
cleanlab/latent_algebra.py - Equalities when noise information is known.
cleanlab/latent_estimation.py - Estimates and fully characterizes all variants of label noise.
cleanlab/noise_generation.py - Generate mathematically valid synthetic noise matrices.
cleanlab/polyplex.py - Characterizes joint distribution of label noise EXACTLY from noisy channel.
cleanlab/pruning.py - Finds the indices of the examples with label errors in a dataset.

Many of these methods have default parameters that won’t be covered here. Check out the method docstrings for full documentation.

Methods to Standardize Research with Noisy Labels

cleanlab supports a number of functions to generate noise for benchmarking and standardization in research. This next example shows how to generate valid, class-conditional, unformly random noisy channel matrices:

# Generate a valid (necessary conditions for learnability are met) noise matrix for any trace > 1
from cleanlab.noise_generation import generate_noise_matrix_from_trace
noise_matrix=generate_noise_matrix_from_trace(
    K=number_of_classes, 
    trace=float_value_greater_than_1_and_leq_K,
    py=prior_of_y_actual_labels_which_is_just_an_array_of_length_K,
    frac_zero_noise_rates=float_from_0_to_1_controlling_sparsity,
)

# Check if a noise matrix is valid (necessary conditions for learnability are met)
from cleanlab.noise_generation import noise_matrix_is_valid
is_valid=noise_matrix_is_valid(noise_matrix, prior_of_y_which_is_just_an_array_of_length_K)

For a given noise matrix, this example shows how to generate noisy labels. Methods can be seeded for reproducibility.

# Generate noisy labels using the noise_marix. Guarantees exact amount of noise in labels.
from cleanlab.noise_generation import generate_noisy_labels
s_noisy_labels = generate_noisy_labels(y_hidden_actual_labels, noise_matrix)

# This package is a full of other useful methods for learning with noisy labels.
# The tutorial stops here, but you don't have to. Inspect method docstrings for full docs.

Estimate the confident joint, the latent noisy channel matrix, P(s | y) and inverse, P(y | s), the latent prior of the unobserved, actual true labels, p(y), and the predicted probabilities.

s denotes a random variable that represents the observed, noisy label and y denotes a random variable representing the hidden, actual labels. Both s and y take any of the m classes as values. The cleanlab package supports different levels of granularity for computation depending on the needs of the user. Because of this, we support multiple alternatives, all no more than a few lines, to estimate these latent distribution arrays, enabling the user to reduce computation time by only computing what they need to compute, as seen in the examples below.

Throughout these examples, you’ll see a variable called confident_joint. The confident joint is an m x m matrix (m is the number of classes) that counts, for every observed, noisy class, the number of examples that confidently belong to every latent, hidden class. It counts the number of examples that we are confident are labeled correctly or incorrectly for every pair of obseved and unobserved classes. The confident joint is an unnormalized estimate of the complete-information latent joint distribution, Ps,y. Most of the methods in the cleanlab package start by first estimating the confident_joint. You can learn more about this in the confident learning paper.

Option 1: Compute the confident joint and predicted probs first. Stop if that’s all you need.

from cleanlab.latent_estimation import estimate_latent
from cleanlab.latent_estimation import estimate_confident_joint_and_cv_pred_proba

# Compute the confident joint and the n x m predicted probabilities matrix (psx),
# for n examples, m classes. Stop here if all you need is the confident joint.
confident_joint, psx = estimate_confident_joint_and_cv_pred_proba(
    X=X_train, 
    s=train_labels_with_errors,
    clf=logreg(), # default, you can use any classifier
)

# Estimate latent distributions: p(y) as est_py, P(s|y) as est_nm, and P(y|s) as est_inv
est_py, est_nm, est_inv = estimate_latent(confident_joint, s=train_labels_with_errors)

Option 2: Estimate the latent distribution matrices in a single line of code.

from cleanlab.latent_estimation import estimate_py_noise_matrices_and_cv_pred_proba
est_py, est_nm, est_inv, confident_joint, psx = estimate_py_noise_matrices_and_cv_pred_proba(
    X=X_train,
    s=train_labels_with_errors,
)

Option 3: Skip computing the predicted probabilities if you already have them.

# Already have psx? (n x m matrix of predicted probabilities)
# For example, you might get them from a pre-trained model (like resnet on ImageNet)
# With the cleanlab package, you estimate directly with psx.
from cleanlab.latent_estimation import estimate_py_and_noise_matrices_from_probabilities
est_py, est_nm, est_inv, confident_joint = estimate_py_and_noise_matrices_from_probabilities(
    s=train_labels_with_errors, 
    psx=psx,
)

Completely characterize label noise in a dataset:

The joint probability distribution of noisy and true labels, P(s,y), completely characterizes label noise with a class-conditional m x m matrix.

from cleanlab.latent_estimation import estimate_joint
joint = compute_confident_joint(
    s=noisy_labels,
    psx=probabilities,
    confident_joint=None,  # Provide if you have it already
)

PU learning with cleanlab:

PU learning is a special case when one of your classes has no error. P stands for the positive class and is assumed to have zero label errors and U stands for unlabeled data, but in practice, we just assume the U class is a noisy negative class that contains some positive examples. Thus, the goal of PU learning is to (1) estimate the proportion of positives in the negative class (see fraction_noise_in_unlabeled_class in the last example), (2) find the errors (see last example), and (3) train on clean data (see first example below). cleanlab does all three, taking into account that there is no label errors in whichever class you specify.

There are two ways to use cleanlab for PU learning. We'll look at each here.

Method 1. If you are using the cleanlab classifier LearningWithNoisyLabels(), and your dataset has exactly two classes (positive = 1, and negative = 0), PU learning is supported directly in cleanlab. You can perform PU learning like this:

from cleanlab.classification import LearningWithNoisyLabels
from sklearn.linear_model import LogisticRegression
# Wrap around any classifier. Yup, you can use sklearn/pyTorch/Tensorflow/FastText/etc.
pu_class = 0 # Should be 0 or 1. Label of class with NO ERRORS. (e.g., P class in PU)
lnl = LearningWithNoisyLabels(clf=LogisticRegression(), pulearning=pu_class)
lnl.fit(X=X_train_data, s=train_noisy_labels)
# Estimate the predictions you would have gotten by training with *no* label errors.
predicted_test_labels = lnl.predict(X_test)

Method 2. However, you might be using a more complicated classifier that doesn't work well with LearningWithNoisyLabels (see this example for CIFAR-10). Or you might have 3 or more classes. Here's how to use cleanlab for PU learning in this situation. To let cleanlab know which class has no error (in standard PU learning, this is the P class), you need to set the threshold for that class to 1 (1 means the probability that the labels of that class are correct is 1, i.e. that class has no error). Here's the code:

import numpy as np
# K is the number of classes in your dataset
# psx are the cross-validated predicted probabilities.
# s is the array/list/iterable of noisy labels
# pu_class is a 0-based integer for the class that has no label errors.
thresholds = np.asarray([np.mean(psx[:, k][s == k]) for k in range(K)])
thresholds[pu_class] = 1.0

Now you can use cleanlab however you were before. Just be sure to pass in this thresholds parameter wherever it applies. For example:

# Uncertainty quantification (characterize the label noise
# by estimating the joint distribution of noisy and true labels)
cj = compute_confident_joint(s, psx, thresholds=thresholds, )
# Now the noise (cj) has been estimated taking into account that some class(es) have no error.
# We can use cj to find label errors like this:
indices_of_label_errors = get_noise_indices(s, psx, confident_joint=cj, )

# In addition to label errors, we can find the fraction of noise in the unlabeled class.
# First we need the inv_noise_matrix which contains P(y|s) (proportion of mislabeling).
_, _, inv_noise_matrix = estimate_latent(confident_joint=cj, s=s, )
# Because inv_noise_matrix contains P(y|s), p (y = anything | s = pu_class) should be 0
# because the prob(true label is something else | example is in pu_class) is 0.
# What's more interesting is p(y = anything | s is not put_class), or in the binary case
# this translates to p(y = pu_class | s = 1 - pu_class) because pu_class is 0 or 1.
# So, to find the fraction_noise_in_unlabeled_class, for binary, you just compute:
fraction_noise_in_unlabeled_class = inv_noise_matrix[pu_class][1 - pu_class]

Now that you have indices_of_label_errors, you can remove those label errors and train on clean data (or only remove some of the label errors and iteratively use confident learning / cleanlab to improve results)

The Polyplex

The key to learning in the presence of label errors is estimating the joint distribution between the actual, hidden labels ‘y’ and the observed, noisy labels ‘s’. Using cleanlab and the theory of confident learning, we can completely characterize the trace of the latent joint distribution, trace(P(s,y)), given p(y), for any fraction of label errors, i.e. for any trace of the noisy channel, trace(P(s|y)).

You can check out how to do this yourself here: 1. Drawing Polyplices 2. Computing Polyplices

KBodolai/cleanlab

So fresh, so `cleanlab`

Find label errors with PyTorch, Tensorflow, MXNet, etc. in 1 line of code.

Learning with noisy labels in 3 lines of code!

Installation

Reproducing Results in confident learning paper

`cleanlab`: State of the Art Learning with Noisy Labels in CIFAR

`cleanlab`: Find Label Errors in ImageNet

`cleanlab`: Find Label Errors in MNIST

`cleanlab` Generality: View performance across 4 distributions and 9 classifiers.

Get started with easy, quick examples.

Use `cleanlab` with any model (Tensorflow, caffe2, PyTorch, etc.)

Want to see a working example? Here’s a compliant PyTorch MNIST CNN class

Documentation by Example

`cleanlab` Core Package Components

Methods to Standardize Research with Noisy Labels

Estimate the confident joint, the latent noisy channel matrix, P(s | y) and inverse, P(y | s), the latent prior of the unobserved, actual true labels, p(y), and the predicted probabilities.

Option 1: Compute the confident joint and predicted probs first. Stop if that’s all you need.

Option 2: Estimate the latent distribution matrices in a single line of code.

Option 3: Skip computing the predicted probabilities if you already have them.

Completely characterize label noise in a dataset:

PU learning with cleanlab:

The Polyplex

License

KBodolai/cleanlab

So fresh, so cleanlab

Find label errors with PyTorch, Tensorflow, MXNet, etc. in 1 line of code.

Learning with noisy labels in 3 lines of code!

Installation

Citations and Related Publications

Reproducing Results in confident learning paper

cleanlab: State of the Art Learning with Noisy Labels in CIFAR

cleanlab: Find Label Errors in ImageNet

cleanlab: Find Label Errors in MNIST

cleanlab Generality: View performance across 4 distributions and 9 classifiers.

Get started with easy, quick examples.

Use cleanlab with any model (Tensorflow, caffe2, PyTorch, etc.)

Want to see a working example? Here’s a compliant PyTorch MNIST CNN class

Documentation by Example

cleanlab Core Package Components

Methods to Standardize Research with Noisy Labels

Estimate the confident joint, the latent noisy channel matrix, P(s | y) and inverse, P(y | s), the latent prior of the unobserved, actual true labels, p(y), and the predicted probabilities.

Option 1: Compute the confident joint and predicted probs first. Stop if that’s all you need.

Option 2: Estimate the latent distribution matrices in a single line of code.

Option 3: Skip computing the predicted probabilities if you already have them.

Completely characterize label noise in a dataset:

PU learning with cleanlab:

The Polyplex

License

So fresh, so `cleanlab`

`cleanlab`: State of the Art Learning with Noisy Labels in CIFAR

`cleanlab`: Find Label Errors in ImageNet

`cleanlab`: Find Label Errors in MNIST

`cleanlab` Generality: View performance across 4 distributions and 9 classifiers.

Use `cleanlab` with any model (Tensorflow, caffe2, PyTorch, etc.)

`cleanlab` Core Package Components