Implement Deep Partition Aggregation

[keykholt]

Description

Implementation of Deep Partition Aggregation (https://arxiv.org/abs/2006.14768) as a classification estimator.

Fixes #1389

Type of change

Please check all relevant options.

• Improvement (non-breaking)
• Bug fix (non-breaking)
• New feature (non-breaking)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

Testing

Please describe the tests that you ran to verify your changes. Consider listing any relevant details of your test configuration.

Tested using the included notebook poisoning_defense_deep_partition_aggregation.ipynb

Test Configuration:

• Ubuntu 18.04.6
• Python v3.8.8
• ART v1.9.0
• TensorFlow 2.4.1

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[lgtm-com]

This pull request introduces 1 alert when merging 8b733c9 into 98627f8 - view on LGTM.com

new alerts:

• 1 for Unused import

[codecov-commenter]

Codecov Report

Merging #1397 (6acda54) into dev_1.9.0 (96ad148) will increase coverage by 0.33%.
The diff coverage is 76.78%.

@@              Coverage Diff              @@
##           dev_1.9.0    #1397      +/-   ##
=============================================
+ Coverage      90.18%   90.52%   +0.33%     
=============================================
  Files            234      235       +1     
  Lines          18993    19049      +56     
  Branches        3349     3365      +16     
=============================================
+ Hits           17129    17244     +115     
+ Misses          1098     1038      -60     
- Partials         766      767       +1     

Impacted Files	Coverage Δ
art/estimators/classification/ensemble.py	94.82% <ø> (ø)
...timators/classification/deep_partition_ensemble.py	76.36% <76.36%> (ø)
art/estimators/classification/__init__.py	100.00% <100.00%> (ø)
art/estimators/classification/tensorflow.py	82.80% <0.00%> (+0.20%)	⬆️
art/attacks/evasion/boundary.py	93.93% <0.00%> (+1.21%)	⬆️
...timators/poison_mitigation/neural_cleanse/keras.py	93.66% <0.00%> (+23.23%)	⬆️
...poison_mitigation/neural_cleanse/neural_cleanse.py	86.48% <0.00%> (+32.43%)	⬆️

[keykholt]

Looks like the final style checks passed after modifying the code according to our discussion

[beat-buesser]

Hi @keykholt Thank you very much for contributing Deep Partition Aggregation!