Samuel Hoffman

Adds a new `average()` helper function which performs a generalized average. Options include: minimum, maximum, harmonic, geometric, arithmetic, and root mean square. This is used in conjunction with `recall_score` and...

Colab for jupyter notebooks in AIF360

2

Make notebooks compatible with Google Colab. This generally involves adding a cell to install the package and download the relevant dataset. See [demo_lime.ipynb](https://github.com/Trusted-AI/AIF360/blob/master/examples/demo_lime.ipynb) for an example. This snippet may be...

Allow different distance metrics in consistency_score

2

- [ ] Add **kwargs to pass through to `NearestNeighbors` to configure the distance metric - [ ] ~Add option to minmax scale X before fitting KNN~ This can happen...

Investigate how to link notebooks in documentation

5

It would be nice if we could link the examples we already have to their respective API doc pages. Can we use [Sphinx-gallery](https://sphinx-gallery.github.io/stable/index.html)? Will it format notebooks properly? How automated...

Download LawSchool dataset directly from SEAPHE

1

http://www.seaphe.org/databases.php This way we can remove the dependency on tempeh. We can essentially copy this file (preserving the copyright notice): https://github.com/microsoft/tempeh/blob/main/tempeh/datasets/seaphe_datasets.py See also [meps_datasets.py](https://github.com/Trusted-AI/AIF360/blob/master/aif360/sklearn/datasets/meps_datasets.py) for another example of downloading/unzipping. Relevant...

Create class/function similar to [RocCurveDisplay](https://scikit-learn.org/stable/modules/generated/sklearn.metrics.RocCurveDisplay.html) which generates diagrams like in [Fig. 3 (Pleiss, et al. 2017)](https://arxiv.org/pdf/1709.02012.pdf). See the [example notebook](https://github.com/Trusted-AI/AIF360/blob/master/examples/sklearn/demo_new_features.ipynb) for a starting point.

Additional parameters for fairlearn reductions

2

- [ ] ~~Pass through `sample_weights` in `fit()`~~ - [ ] A few optional args are missing from our initializer (e.g. `grid_offset`) - [ ] Optional `random_state` arg for `__init__()`...

* add `priv_group` option to binarize groups * make `pos_label` implicit if constraint is `'weighted'`

[AdversarialDebiasing](https://github.com/Trusted-AI/AIF360/blob/master/aif360/sklearn/inprocessing/adversarial_debiasing.py) currently uses TFv1 code which is pretty out of date and doesn't allow for custom models. Would be nice to update this. Since we already use PyTorch elsewhere, it...

Add User Guide to docs

2

- [ ] Datasets - [ ] Detectors - [ ] In-processing - [ ] Metrics - [ ] Post-processing - [ ] Pre-processing