Harnessing value from data science in business: ensuring explainability and fairness of solutions
arXiv:2108.07714
Abstract
The paper introduces concepts of fairness and explainability (XAI) in artificial intelligence, oriented to solve a sophisticated business problems. For fairness, the authors discuss the bias-inducing specifics, as well as relevant mitigation methods, concluding with a set of recipes for introducing fairness in data-driven organizations. Additionally, for XAI, the authors audit specific algorithms paired with demonstrational business use-cases, discuss a plethora of techniques of explanations quality quantification and provide an overview of future research avenues.
References in corpus (18)
- Auto-Encoding Variational Bayes
- A Unified Approach to Interpreting Model Predictions
- UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction
- Learning Important Features Through Propagating Activation Differences
- SmoothGrad: removing noise by adding noise
- Deep Convolutional Inverse Graphics Network
- Attention is not Explanation
- Deep k-Nearest Neighbors: Towards Confident, Interpretable and Robust Deep Learning
- Deep and Confident Prediction for Time Series at Uber
- On Fairness and Calibration
- The (Un)reliability of saliency methods
- Deep Kalman Filters
- The Bayesian Case Model: A Generative Approach for Case-Based Reasoning and Prototype Classification
- NBDT: Neural-Backed Decision Trees
- Measurable Counterfactual Local Explanations for Any Classifier
- Learning unbiased features
- Hybrid Predictive Model: When an Interpretable Model Collaborates with a Black-box Model
- Regularizing Black-box Models for Improved Interpretability (HILL 2019 Version)