---

Alibi is a Python library aimed at machine learning model inspection and interpretation. The focus of the library is to provide high-quality implementations of black-box, white-box, local and global explanation methods for classification and regression models.

• Documentation

If you're interested in outlier detection, concept drift or adversarial instance detection, check out our sister project alibi-detect.

Anchor explanations for images	Integrated Gradients for text
Counterfactual examples	Accumulated Local Effects

Table of Contents

• Installation and Usage
• Supported Methods
  • Model Explanations
  • Model Confidence
  • Prototypes
  • References and Examples
• Citations

Installation and Usage

Alibi can be installed from:

• PyPI or GitHub source (with pip)
• Anaconda (with conda/mamba)

With pip

• Alibi can be installed from PyPI:

  pip install alibi

• Alternatively, the development version can be installed:

  pip install git+https://github.com/SeldonIO/alibi.git 

• To take advantage of distributed computation of explanations, install alibi with ray:

  pip install alibi[ray]

• For SHAP support, install alibi as follows:

  pip install alibi[shap]

With conda

To install from conda-forge it is recommended to use mamba, which can be installed to the base conda enviroment with:

conda install mamba -n base -c conda-forge

• For the standard Alibi install:

  mamba install -c conda-forge alibi

• For distributed computing support:

  mamba install -c conda-forge alibi ray

• For SHAP support:

  mamba install -c conda-forge alibi shap

Usage

The alibi explanation API takes inspiration from scikit-learn, consisting of distinct initialize, fit and explain steps. We will use the AnchorTabular explainer to illustrate the API:

from alibi.explainers import AnchorTabular

# initialize and fit explainer by passing a prediction function and any other required arguments
explainer = AnchorTabular(predict_fn, feature_names=feature_names, category_map=category_map)
explainer.fit(X_train)

# explain an instance
explanation = explainer.explain(x)

The explanation returned is an Explanation object with attributes meta and data. meta is a dictionary containing the explainer metadata and any hyperparameters and data is a dictionary containing everything related to the computed explanation. For example, for the Anchor algorithm the explanation can be accessed via explanation.data['anchor'] (or explanation.anchor). The exact details of available fields varies from method to method so we encourage the reader to become familiar with the types of methods supported.

Supported Methods

The following tables summarize the possible use cases for each method.

Model Explanations

Method	Models	Explanations	Classification	Regression	Tabular	Text	Images	Categorical features	Train set required	Distributed
ALE	BB	global	✔	✔	✔
Partial Dependence	BB WB	global	✔	✔	✔			✔
PD Variance	BB WB	global	✔	✔	✔			✔
Permutation Importance	BB	global	✔	✔	✔			✔
Anchors	BB	local	✔		✔	✔	✔	✔	For Tabular
CEM	BB* TF/Keras	local	✔		✔		✔		Optional
Counterfactuals	BB* TF/Keras	local	✔		✔		✔		No
Prototype Counterfactuals	BB* TF/Keras	local	✔		✔		✔	✔	Optional
Counterfactuals with RL	BB	local	✔		✔		✔	✔	✔
Integrated Gradients	TF/Keras	local	✔	✔	✔	✔	✔	✔	Optional
Kernel SHAP	BB	local global	✔	✔	✔			✔	✔	✔
Tree SHAP	WB	local global	✔	✔	✔			✔	Optional
Similarity explanations	WB	local	✔	✔	✔	✔	✔	✔	✔

Model Confidence

These algorithms provide instance-specific scores measuring the model confidence for making a particular prediction.

Method	Models	Classification	Regression	Tabular	Text	Images	Categorical Features	Train set required
Trust Scores	BB	✔		✔	✔(1)	✔(2)		Yes
Linearity Measure	BB	✔	✔	✔		✔		Optional

Key:

• BB - black-box (only require a prediction function)
• BB* - black-box but assume model is differentiable
• WB - requires white-box model access. There may be limitations on models supported
• TF/Keras - TensorFlow models via the Keras API
• Local - instance specific explanation, why was this prediction made?
• Global - explains the model with respect to a set of instances
• (1) - depending on model
• (2) - may require dimensionality reduction

Prototypes

These algorithms provide a distilled view of the dataset and help construct a 1-KNN interpretable classifier.

Method	Classification	Regression	Tabular	Text	Images	Categorical Features	Train set labels
ProtoSelect	✔		✔	✔	✔	✔	Optional

References and Examples

• Accumulated Local Effects (ALE, Apley and Zhu, 2016)

  • Documentation
  • Examples: California housing dataset, Iris dataset

• Partial Dependence (J.H. Friedman, 2001)

  • Documentation
  • Examples: Bike rental

• Partial Dependence Variance(Greenwell et al., 2018)

  • Documentation
  • Examples: Friedman’s regression problem

• Permutation Importance(Breiman, 2001; Fisher et al., 2018)

  • Documentation
  • Examples: Who's Going to Leave Next?

• Anchor explanations (Ribeiro et al., 2018)

  • Documentation
  • Examples: income prediction, Iris dataset, movie sentiment classification, ImageNet, fashion MNIST

• Contrastive Explanation Method (CEM, Dhurandhar et al., 2018)

  • Documentation
  • Examples: MNIST, Iris dataset

• Counterfactual Explanations (extension of Wachter et al., 2017)

  • Documentation
  • Examples: MNIST

• Counterfactual Explanations Guided by Prototypes (Van Looveren and Klaise, 2019)

  • Documentation
  • Examples: MNIST, California housing dataset, Adult income (one-hot), Adult income (ordinal)

• Model-agnostic Counterfactual Explanations via RL(Samoilescu et al., 2021)

  • Documentation
  • Examples: MNIST, Adult income

• Integrated Gradients (Sundararajan et al., 2017)

  • Documentation,
  • Examples: MNIST example, Imagenet example, IMDB example.

• Kernel Shapley Additive Explanations (Lundberg et al., 2017)

  • Documentation
  • Examples: SVM with continuous data, multinomial logistic regression with continous data, handling categorical variables

• Tree Shapley Additive Explanations (Lundberg et al., 2020)

  • Documentation
  • Examples: Interventional (adult income, xgboost), Path-dependent (adult income, xgboost)

• Trust Scores (Jiang et al., 2018)

  • Documentation
  • Examples: MNIST, Iris dataset

• Linearity Measure

  • Documentation
  • Examples: Iris dataset, fashion MNIST

• ProtoSelect

  • Documentation
  • Examples: Adult Census & CIFAR10

• Similarity explanations

  • Documentation
  • Examples: 20 news groups dataset, ImageNet dataset, MNIST dataset

Citations

If you use alibi in your research, please consider citing it.

BibTeX entry:

@article{JMLR:v22:21-0017,
  author  = {Janis Klaise and Arnaud Van Looveren and Giovanni Vacanti and Alexandru Coca},
  title   = {Alibi Explain: Algorithms for Explaining Machine Learning Models},
  journal = {Journal of Machine Learning Research},
  year    = {2021},
  volume  = {22},
  number  = {181},
  pages   = {1-7},
  url     = {http://jmlr.org/papers/v22/21-0017.html}
}