AIF360 v0.6.1 Release Notes

Highlights

• New detector: FACTS/FACTS_bias_scan

What's Changed

• Fairness Aware Counterfactuals for Subgroups by @phantom-duck in #457
• Avoiding uses of np.float because that has been removed from the latest numpy. by @hirzel in #521
• Fixed Example Google Style Python Docstrings URL by @ShorthillsAI in #522
• Add FACTS import warnings by @hoffmansc in #524
• Bump version -> 0.6.1 by @hoffmansc in #525

New Contributors

• @phantom-duck made their first contribution in #457
• @hirzel made their first contribution in #521
• @ShorthillsAI made their first contribution in #522

Full Changelog: v0.6.0...v0.6.1

AIF360 v0.6.0 Release Notes

Highlights

• New algorithms:
  • SenSeI/SenSR
  • DeterministicReranking
• New metric:
  • ot_distance

Backwards-Incompatible Changes

• Dropped support for bias_scan from aif360.metrics/aif360.sklearn.metrics
• Minor changes to MEPS files

What's Changed

• Add python 3.10 for testing in ci.yml by @hakimamarullah in #368
• Allow binder to automatically build environment by @hoffmansc in #395
• Update to demo_lime notebook to run out of the box (and in Google collab) by @anupamamurthi in #396
• Fix error in metric_json_explainer.consistency() by @hoffmansc in #400
• Solving issue #381, to run demo_new_features notebook in Colab by @ivesulca in #402
• logic error in disparate_impact and statistical_parity_difference #292 by @sreeja-g in #407
• Modify example notebooks to work in Google colab #378 by @Yashaswini-Viswanath in #405
• Solving second part of issue #381 by @ivesulca in #409
• Get notebooks in examples/sklearn/ to work in Google colab (Part 2/3) [#380] by @dharmod in #403
• average predictive value difference metric implementation #376 by @sreeja-g in #410
• Add code coverage checks with pytest-cov by @aitorres in #412
• Initial inFairness algorithms (SenSeI/SenSR) by @hoffmansc in #340
• Remove turtle module import by @gkumbhat in #415
• Contribution Guide by @hoffmansc in #427
• Bump versions by @hoffmansc in #458
• Fix typo in GridSearchReduction by @haas-christian in #452
• Add a bias detector based on optimal transport by @jmarecek in #434
• Issue 265 privileged class bank dataset by @joosjegoedhart in #449
• Fix typo. by @gowriaddepalli in #470
• Added methods for equalized odds difference by @divyagaddipati in #477
• Update data_preproc_functions.py to solve conversion to float issue by @baraldian in #498
• Add methods for dealing with fairness in rankings by @andrewklayk in #461
• Update homepage URL by @hoffmansc in #503
• Remove deprecated bias scan metrics by @hoffmansc in #504
• Replace deprecated if_delegate_has_method with available_if by @hoffmansc in #511
• Fix tests failing due to int columns by @hoffmansc in #513
• Change source for Law School GPA dataset by @hoffmansc in #510
• Install R dependencies by @hoffmansc in #514
• Update sphinx requirement by @hoffmansc in #512
• Bump jinja2 from 3.0.3 to 3.1.3 by @dependabot in #507
• add .readthedocs.yaml by @hoffmansc in #516
• Fix .readthedocs.yaml and bump version by @hoffmansc in #517
• Fix sphinx_rtd_theme by @hoffmansc in #518
• Rename master -> main by @hoffmansc in #515
• Remove requests dependency by @hoffmansc in #519
• Include fairadapt.R in package by @hoffmansc in #520

New Contributors

• @hakimamarullah made their first contribution in #368
• @anupamamurthi made their first contribution in #396
• @ivesulca made their first contribution in #402
• @sreeja-g made their first contribution in #407
• @Yashaswini-Viswanath made their first contribution in #405
• @dharmod made their first contribution in #403
• @aitorres made their first contribution in #412
• @gkumbhat made their first contribution in #415
• @haas-christian made their first contribution in #452
• @jmarecek made their first contribution in #434
• @joosjegoedhart made their first contribution in #449
• @gowriaddepalli made their first contribution in #470
• @divyagaddipati made their first contribution in #477
• @baraldian made their first contribution in #498
• @andrewklayk made their first contribution in #461

Full Changelog: v0.5.0...v0.6.0

AIF360 v0.5.0 Release Notes

Highlights

• New algorithms:
  • FairAdapt
• New metrics:
  • MDSS
  • class_imbalance, kl_divergence, conditional_demographic_disparity
  • intersection and one_vs_rest meta-metrics
• sklearn-compatible ports:
  • differential fairness metrics
  • MEPS, COMPAS violent
  • RejectOptionClassification, LearnedFairRepresentations

New Features/Improvements

• Multidimensional subset scanning (MDSS) for bias in classifiers by @Viktour19 in #238
• Update component.yaml to kfp v2 sdk by @yhwang in #259
• Fairadapt inclusion in AIF360 by @dplecko in #257
• Added a tutorial for advertising data by @barvek in #310
• More sklearn-compatible algorithms by @hoffmansc in #318
• Dataset Improvements by @hoffmansc in #278
  • array of sample-wise protected attributes may now be passed in prot_attr instead of an index label
• Method of the month (July) by @hoffmansc in #324
• sklearn-compat additions by @mnagired in #322
  • add predict_proba to RejectOptionClassifier
• More sklearn-compatible metrics by @hoffmansc in #290
  • smoothed_edf, df_bias_amplification
  • class_imbalance, kl_divergence, conditional_demographic_disparity
  • intersection, one_vs_rest

Backwards-Incompatible Changes

• Add detectors api by @Adebayo-Oshingbesan in #305
  • version of bias_scan in aif360.metrics to be deprecated next release

Fixes

• Fixed computation of coefficient of variation in classification_metrics by @plankington in #288
• Fix exponential gradient reduction without protected attribute (#267) by @jdnklau in #268
• Remove caches due to excessive memory use by @Adebayo-Oshingbesan in #317
• fix rpy2 crash bug by @hoffmansc in #313
• Fix pipelining bug in fairlearn algorithms by @hoffmansc in #323
• Optional tempeh, conditional imports by @DanielRyszkaIBM in #338
• Restricting AdversarialDebiasing's trainable variables to current scope by @mfeffer in #255
• Increasing max_iter to 1000 for LogisticRegression used in PrejudiceRemover by @mfeffer in #254

New Contributors

• @Viktour19 made their first contribution in #238
• @jdnklau made their first contribution in #268
• @yhwang made their first contribution in #259
• @dplecko made their first contribution in #257
• @plankington made their first contribution in #288
• @Adebayo-Oshingbesan made their first contribution in #305
• @barvek made their first contribution in #310
• @milevavantuyl made their first contribution in #309
• @josue-rodriguez made their first contribution in #315
• @DanielRyszkaIBM made their first contribution in #338
• @mnagired made their first contribution in #322
• @mfeffer made their first contribution in #255

Full Changelog: v0.4.0...v0.5.0

AIF360 v0.4.0 Release Notes

This is a major release containing a number of new features, improvements, and bugfixes.

Highlights

• TensorFlow 2, Python 3.8 now supported
• New algorithms:
  • Exponentiated Gradient Reduction
  • Grid Search Reduction
• New dataset:
  • Law school GPA

New Features/Improvements

• Python 3.8 and TensorFlow 2 (via compat.v1) support added (#230)
• Algorithms from fairlearn added (#215):
  • Exponentiated Gradient Reduction and Grid Search Reduction
  • Support for regression datasets
  • Law school GPA dataset added
• MetaFairClassifier code cleaned and sped up (#196)
• removed numba dependency (#187)
• maxiter and maxfun arguments in LFR fit() (#184)

Backwards-Incompatible Changes

• Removed support for Python 3.5

Fixes

• Fix bug where scores in a single-row dataset was getting squeezed (#193)
• Typo in consistency_score documentation (#195)
• Lime notebook license issue (#191)

New Contributors

@baba-mpe, @SSaishruthi, @leenamurgai, @synapticarbors, @sohiniu, @yangky11

AIF360 v0.3.0 Release Notes

This is a major release containing a number of new features, improvements, and bugfixes.

Highlights

• scikit-learn compatible API for certain algorithms, metrics, and datasets
• Documentation layout was revamped to make it easier to navigate.
• New algorithm:
  • Fairness Gerrymandering (Kearns, et al., 2018)
• New metrics:
  • Differential Fairness (Foulds, et al., 2018)
  • Rich Subgroup Fairness (Kearns, et al., 2018)

New Features/Improvements

• Optional dependencies may now be installed using the setuptools "extras" option: e.g., pip install 'aif360[LFR,AdversarialDebiasing]' or pip install 'aif360[all]'
• Added support for integrations with MLOps (Kubeflow and NiFi) and examples
• Added scores output to AdversarialDebiasing.predict() (#139)
• Added a subset() method to StructuredDataset (#140)
• Added new MulticlassLabelDataset to support basic multiclass problems (#165)

• scikit-learn compatibility (#134)

  • EXPERIMENTAL: incomplete, contributions welcome
  • 4 datasets (Adult, German, Bank, Compas) in DataFrame format with protected attributes in the index
    • Automatically downloads from openml.org
  • 6 group fairness metrics as functions (statistical_parity_difference, disparate_impact_ratio, equal_opportunity_difference, average_odds_difference, average_odds_error, between_group_generalized_entropy_error)
  • 2 individual fairness metrics as functions (generalized_entropy_index and its variants, consistency_score)
  • 5 additional metrics as functions (specificity_score, base_rate, selection_rate, generalized_fpr, generalized_fnr)
  • make_scorer function to wrap metrics for use in sklearn cross-validation functions (#174, #178)
  • 3 algorithms (Reweighing, AdversarialDebiasing, CalibratedEqualizedOdds)

Fixes

• Fixed deprecation warning/NotImplementedError in StandardDataset (#115)
• Fixed age threshold in GermanDataset (#129 and #137)
• Corrected privileged/unprivileged attribute values for COMPAS dataset in some demos (#138)
• Fixed base rate computation in EqOddsPostprocessing (#170)
• Improved warning messages when missing optional packages (#170)
• Multiple documentation fixes (#114, #124, #153, #155, #157, #158, #159, #170)

New Contributors

@autoih, @romeokienzler, @jimbudarz, @stephanNorsten, @sethneel, @imolloy, @guillemarsan, @gdequeiroz, @chajath, @bhavyaghai, @Tomcli, @swapna-somineni, @chkoar, @motapaolla

AIF360 v0.3.0rc0 Release Notes

This is a major release containing a number of new features, improvements, and bugfixes.

Highlights

• scikit-learn compatible API for certain algorithms, metrics, and datasets
• Documentation layout was revamped to make it easier to navigate.
• New algorithm:
  • Fairness Gerrymandering (Kearns, et al., 2018)
• New metrics:
  • Differential Fairness (Foulds, et al., 2018)
  • Rich Subgroup Fairness (Kearns, et al., 2018)

New Features/Improvements

• Optional dependencies may now be installed using the setuptools "extras" option: e.g., pip install 'aif360[LFR,AdversarialDebiasing]' or pip install 'aif360[all]'
• Added support for integrations with MLOps (Kubeflow and NiFi) and examples
• Added scores output to AdversarialDebiasing.predict() (#139)
• Added a subset() method to StructuredDataset (#140)

• scikit-learn compatibility (#134)

  • EXPERIMENTAL: incomplete, contributions welcome
  • 4 datasets (Adult, German, Bank, Compas) in DataFrame format with protected attributes in the index
    • Automatically downloads from openml.org
  • 6 group fairness metrics as functions (statistical_parity_difference, disparate_impact_ratio, equal_opportunity_difference, average_odds_difference, average_odds_error, between_group_generalized_entropy_error)
  • 2 individual fairness metrics as functions (generalized_entropy_index and its variants, consistency_score)
  • 5 additional metrics as functions (specificity_score, base_rate, selection_rate, generalized_fpr, generalized_fnr)
  • 3 algorithms (Reweighing, AdversarialDebiasing, CalibratedEqualizedOdds)

Fixes

• Fixed deprecation warning/NotImplementedError in StandardDataset (#115)
• Fixed age threshold in GermanDataset (#129 and #137)
• Corrected privileged/unprivileged attribute values for COMPAS dataset in some demos (#138)
• Multiple documentation fixes (#114, #124, #153, #155, #157, #158, #159)

New Contributors

@autoih, @romeokienzler, @jimbudarz, @stephanNorsten, @sethneel, @imolloy, @guillemarsan, @gdequeiroz, @chajath, @bhavyaghai, @Tomcli

AIF360 v0.2.3 Release Notes

Fixes

• Fixed fit_predict arguments in RejectOptionClassification (#111)
• Removed Orange3 from requirements (#113)

AIF360 v0.2.2 Release Notes

Fixes

• Removed Gender Classification tutorial (see #101 for details and discussion)
• Bug fix in Optimized Preprocessing to check for optimality correctly

AIF360 v0.2.1 Release Notes

Backwards-Incompatible Changes

• Deprecated support for Python 2.7

Fixes

• See issues #80, #83
• Also PRs #86, #90

AIF360 v0.2.0 Release Notes

Highlights

New Algorithm:

• Meta-Algorithm for Fair Classification (Celis et al.. 2018)

New Features/Improvements

• Added download script for MEPS data
• Added ability to choose protected attribute for DisparateImpactRemover
• Updated OptimPreproc to use the latest version of cvxpy
• Added a threshold value to update labels from predicted scores in CalibratedEqOddsPostprocessing
• New scores_names arg in StructuredDataset allows for easier importing of predictions run elsewhere
• tutorial_gender_classification notebook now uses skimage instead of cv2
• aif360.__version__ now returns the correct version string

Fixes

• Changed Credit Scoring Tutorial to use Reweighing; added new demo using AdversarialDebiasing on Adult Dataset
• Removed dependency on subprocess.run in PrejudiceRemover for Python 2.7 compatibility
• Fixed bug where categorical_features would not take into account features_to_drop in StandardDataset

New Contributors

@ckadner, @cclauss, @vijaykeswani, @ffosilva, @kant, @adrinjalali, @mariaborbones