/fact-ai

Repository for course Fairness Accountability, Confidentiality and Transparency given by the University of Amsterdam

Primary LanguageJupyter Notebook

General instructions

Please be aware that we have two different versions of the code belonging to this project:

  1. The originally provided code by Plumb. et al, upgraded to TensorFlow 2.x
    1. For the (by Plumb et al.) pre-trained models that are simply evaluated, head to ELDR-TF2.x_(pre_trained_models) for further instructions
    2. For the (by us) re-trained models, using the upgraded existing code, head to ELDR-TF2.x_(newly_trained_models) for further instructions
  2. Our new, from-scratch, implementation which includes additional dimensionality reduction algorithms and datasets - head to ELDR-NEW

Dataset information

Seeds Glass Wine Heart Iris Housing RNA
Data Set Characteristics Multivariate Multivariate Multivariate Multivariate Multivariate Multivariate Multivariate
Attribute Characteristics Real Real Integer, Real Categorical, Integer, Real Real Real Real
Associated Tasks Classification, Clustering Classification Classification Classification Classification Classification Classification
Number of Instances 210 214 178 303 150 506 25000
Number of Attributes 7 13 10 75* 4 13 13166
Missing Values N/A No No Yes No No No
Area Life Physical Physical Life Life Physical Life

*only 14 attributes used

Source

Housing: https://archive.ics.uci.edu/ml/machine-learning-databases/housing/

Iris: https://archive.ics.uci.edu/ml/datasets/iris

Heart: https://archive.ics.uci.edu/ml/datasets/heart+disease

Seeds: https://archive.ics.uci.edu/ml/datasets/seeds

Glass: https://archive.ics.uci.edu/ml/datasets/glass+identification

Wine: https://archive.ics.uci.edu/ml/datasets/wine