SLISEMAP: supervised dimensionality reduction through local explanations (2022-01-12T00:00:00.000000Z)

TL;DR

This work proposes a new supervised manifold visualisation method, slisemap, that simultaneously finds local explanations for all data items and builds a (typically) two-dimensional global visualisation of the black box model such that data items with similar local explanations are projected nearby.

Abstract

Existing methods for explaining black box learning models often focus on building local explanations of the models’ behaviour for particular data items. It is possible to create global explanations for all data items, but these explanations generally have low fidelity for complex black box models. We propose a new supervised manifold visualisation method, slisemap , that simultaneously finds local explanations for all data items and builds a (typically) two-dimensional global visualisation of the black box model such that data items with similar local explanations are projected nearby. We provide a mathematical derivation of our problem and an open source implementation implemented using the GPU-optimised PyTorch library. We compare slisemap to multiple popular dimensionality reduction methods and find that slisemap is able to utilise labelled data to create embeddings with consistent local white box models. We also compare slisemap to other model-agnostic local explanation methods and show that slisemap provides comparable explanations and that the visualisations can give a broader understanding of black box regression and classification models.

Authors

Anton Björklund

1 papers

J. Mäkelä

1 papers

Kai Puolamäki

1 papers

TL;DR

Abstract

Authors

References66 items

Robust regression via error tolerance

Detecting virtual concept drift of regressors without ground truth values

Conditional t-SNE: more informative t-SNE embeddings

A review of UMAP in population genetics

The elements of statistical learning: data mining, inference, and prediction, 2nd Edition

PyTorch: An Imperative Style, High-Performance Deep Learning Library

Sparse Robust Regression for Explaining Classifiers

SenseBERT: Driving Some Sense into BERT

Machine Learning With Python

A New Supervised t-SNE with Dissimilarity Measure for Effective Data Visualization and Classification

Unmasking Clever Hans predictors and assessing what machines really learn

The art of using t-SNE for single-cell transcriptomics

UMAP: Uniform Manifold Approximation and Projection

Defining Locality for Surrogates in Post-hoc Interpretablity

Local Rule-Based Explanations of Black Box Decision Systems

Anchors: High-Precision Model-Agnostic Explanations

Dissecting stellar chemical abundance space with t-SNE

UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction

A Survey of Methods for Explaining Black Box Models

All Models are Wrong, but Many are Useful: Learning a Variable's Importance by Studying an Entire Class of Prediction Models Simultaneously

Interpreting Classifiers through Attribute Interactions in Datasets

A Unified Approach to Interpreting Model Predictions

Interpretable Explanations of Black Boxes by Meaningful Perturbation

Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization

European Union Regulations on Algorithmic Decision-Making and a "Right to Explanation"

Algorithmic Transparency via Quantitative Input Influence: Theory and Experiments with Learning Systems

Auditing black-box models for indirect influence

“Why Should I Trust You?”: Explaining the Predictions of Any Classifier

A peek into the black box: exploring classifiers by randomization

OpenML: networked science in machine learning

Linear dimensionality reduction: survey, insights, and generalizations

Searching for exotic particles in high-energy physics with deep learning

Peeking Inside the Black Box: Visualizing Statistical Learning With Plots of Individual Conditional Expectation

Local Linear Regression on Manifolds and Its Geometric Interpretation

Scikit-learn: Machine Learning in Python

How to Explain Individual Classification Decisions

The Elements of Statistical Learning: Data Mining, Inference, and Prediction, Second Edition by Trevor Hastie, Robert Tibshirani, Jerome Friedman

MLLE: Modified Locally Linear Embedding Using Multiple Weights

On Divergences and Informations in Statistics and Information Theory

The Elements of Statistical Learning: Data Mining, Inference, and Prediction

Laplacian Eigenmaps for Dimensionality Reduction and Data Representation

A global geometric framework for nonlinear dimensionality reduction.

Nonlinear dimensionality reduction by locally linear embedding.

Local Linear Model Trees (LOLIMOT) Toolbox for Nonlinear System Identification

Review of the Development of Multidimensional Scaling Methods

Updating Quasi-Newton Matrices With Limited Storage

Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis

SLISEMAP: Combining Supervised Dimensionality Reduction with Local Explanations

2022a) SLISEMAP. URL https://github.com/ edahelsinki/slisemap

Finnish grid and cloud

Explaining t-SNE Embeddings Locally by Adapting LIME

The Right to an Explanation

Explainable AI: Interpreting, Explaining and Visualizing Deep Learning

Interpretable Machine Learning: A Guide for Making Black Box Models Interpretable

A 3D Playground for t-SNE With explainable classification

Inference and Prediction

2018) in the methods

Accelerating t-SNE using tree-based algorithms

Visualizing Data using t-SNE

The Elements of Statistical Learning: Data Mining, Inference, and Prediction

Gradient-based learning applied to document recognition

Spam!

Regression Shrinkage and Selection via the Lasso

A General Class of Coefficients of Divergence of One Distribution from Another

Notes on the n-Person Game — II: The Value of an n-Person Game

Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations

Field of Study

Journal Information

Name

Page

Volume

Venue Information

Name

Type

URL

Alternate Names