DrWhy.AI - the collection of tools for Visual Exploration, Explanation and Debugging of Predictive Models

It takes a village to raise a child model.

The way how we do predictive modeling is very ineffective. We spend way too much time on manual time consuming and easy to automate activities like data cleaning and exploration, crisp modeling, model validation. Instead of focusing on model understanding, productisation and communication.

Here we gather tools that can be use to make out work more efficient through the whole model lifecycle. The unified grammar beyond DrWhy.AI universe is described in the Predictive Models: Visual Exploration, Explanation and Debugging book.

Lifecycle for Predictive Models

Tools that are usefull during the model lifetime. stands for our internal tools.

1. Data preparation

• dataMaid; A Suite of Checks for Identification of Potential Errors in a Data Frame as Part of the Data Screening Process
• inspectdf; A collection of utilities for columnwise summary, comparison and visualisation of data frames.
• ggplot2; System for declaratively creating graphics, based on The Grammar of Graphics.

2. Data understanding

• Model Agnostic Variable Importance Scores. Surrogate learning = Train an elastic model and measure feature importance in such model. See DALEX, Model Class Reliance MCR
• vip Variable importance plots
• SAFE Surrogate learning = Train an elastic model and extract feature transformations.
• xspliner Using surrogate black-boxes to train interpretable spline based additive models
• factorMerger Set of tools for factors merging paper
• ingredients Set of tools for model level feature effects and feature importance.

4. Model assembly

• mlr Machine Learning in R paper
• caret Classification And Regression Training paper

5. Model audit

• auditor model verification, validation, and error analysis vigniette
• DALEX Descriptive mAchine Learning EXplanations
• iml; interpretable machine learning R package
• randomForestExplainer A set of tools to understand what is happening inside a Random Forest
• survxai Explanations for survival models paper

6. Model delivery

• breakDown, pyBreakDown and breakDown2 Model Agnostic Explainers for Individual Predictions (with interactions)
• ceterisParibus, pyCeterisParibus, ceterisParibusD3 and ceterisParibus2 Ceteris Paribus Plots (What-If plots) for explanations of a single observation
• localModel and live LIME-like explanations with interpretable features based on Ceteris Paribus curves.
• lime; Local Interpretable Model-Agnostic Explanations (R port of original Python package)
• shapper An R wrapper of SHAP python library
• modelDown modelDown generates a website with HTML summaries for predictive models
• drifter Concept Drift and Concept Shift Detection for Predictive Models
• archivist A set of tools for datasets and plots archiving paper

DrWhy.AI indicator panel

Active development and maintenance

These packages are actively developed and have active maintainer.

• archivist (maintainer: pbiecek)
• DALEX (maintainer: pbiecek)
• auditor (maintainer: agosiewska)
• survxai (maintainer: agosiewska)
• shapper (maintainer: agosiewska)
• iBreakDown (maintainer: pbiecek)
• ingredients (maintainer: pbiecek)
• drifter (maintainer: pbiecek)
• localModel (maintainer: mstaniak)
• modelDown (maintainer: magda-tatarynowicz)

Experimental pre-seed phase (under active development)

• vivo (maintainer: kozaka93)
• EIX (maintainer: ekarbowiak)
• xspliner (maintainer: krystian8207)
• pyDALEX (maintainer: magda-tatarynowicz)
• SAFE for R (maintainer: AnnaGierlak)
• SAFE for Python (maintainer: olagacek)
• pyCeterisParibus (maintainer: kmichael08)
• ceterisParibusD3 (maintainer: flaminka)

Experimental or without maintenance (looking for maintainer!!!)

These packages contain useful features, are still in use but we are looking for an active maintainer.

• randomForestExplainer
• factorMerger
• cr17
• MLGenSig
• modelDown
• pyBreakDown

In the sunset phase, without maintenance

Key features from these packages are copied to another packages.

• ceterisParibus (development moved to ingredients)
• ceterisParibus2 (development moved to ingredients)
• DALEX2 (development moved to DALEX)
• breakDown (development moved to iBreakDown)
• live (development moved to localModel)

Family of Model Explainers

Architecture of DrWhy

DrWhy works on fully trained predictive models. Models can be created with any tool.

DrWhy uses DALEX2 package to wrap model with additional matadata required for explanations, like validation data, predict function etc.

Explainers for predictive models can be created with model agnostic or model specific functions implemented in various packages.