Steven Drucker Microsoft Research A Design Probe to Understand How Data Scientists Understand Machine Learning Models Gamut CHI 2019 Andrew Head UC Berkeley Fred Hohman @fredhohman Georgia Tech Rob DeLine Microsoft Research Rich Caruana Microsoft Research
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Steven Drucker Microsoft Research
A Design Probe to Understand How Data Scientists Understand Machine Learning Models
Gamut
CHI 2019
Andrew Head UC BerkeleyFred Hohman
@fredhohman Georgia Tech Rob DeLine
Microsoft Research
Rich Caruana Microsoft Research
!2
While building and deploying ML models is now an increasingly common practice, interpreting models is not.
!3
What is interpretability?
What is interpretability?
Human understanding of a system’s…
What is interpretability?
Human understanding of a system’s…
internals e.g., components [Gilpin, 2018]
operations e.g., math [Biran, 2017]
data mapping e.g., input to output [Montavon, 2017]
representation in an explanation [Ribeiro, 2016]
What is interpretability?
Human understanding of a system’s…
internals e.g., components [Gilpin, 2018]
operations e.g., math [Biran, 2017]
data mapping e.g., input to output [Montavon, 2017]
representation in an explanation [Ribeiro, 2016]
No formal, agreed upon definition [Lipton, 2016]
!5
GDPR (General Data Protection Regulation)
!5
Chapter 3 Section 4
“Automated individual decision-making, including profiling”
GDPR (General Data Protection Regulation)
Article 22
Right to explanation⚖
!5
Chapter 3 Section 4
“Automated individual decision-making, including profiling”
GDPR (General Data Protection Regulation)
Article 22
!6
Capabilities of interpretability
Design Probe embodying capabilities
Evaluation & Investigation of probe & emerging practice of interpretability w/ real users
Gamut Contributions
1.
2.
3.
Can we operationalize interpretability?
!7
Contribution 1: Interpretability Capabilities
Can we operationalize interpretability?
Formative research with professional data scientists @• 4 senior ML researchers• 5 ML practitioners
!7
Contribution 1: Interpretability Capabilities
Can we operationalize interpretability?
Formative research with professional data scientists @• 4 senior ML researchers• 5 ML practitioners
Prompt: In a perfect world, given a machine learning model, what questions would you ask it to help you interpret both the model and its predictions?
!7
Contribution 1: Interpretability Capabilities
Explainable ML Interface QuestionsFrom formative research
Explainable ML Interface QuestionsFrom formative research
Why does this house cost that much?
Explainable ML Interface QuestionsFrom formative research
Why does this house cost that much?
What is the difference between these two?
$250,000 $250,000
Explainable ML Interface QuestionsFrom formative research
Why does this house cost that much?
What is the difference between these two?
$250,000 $350,000
Explainable ML Interface QuestionsFrom formative research
Why does this house cost that much?
What is the difference between these two?
What if I added…
$ + $
=1 bedroom
Explainable ML Interface QuestionsFrom formative research
Why does this house cost that much?
What is the difference between these two?
What if I added…
What are similar homes?
Explainable ML Interface QuestionsFrom formative research
What if I added…
What are similar homes? ML
Why does this house cost that much?
What is the difference between these two?
Where is it wrong?
Explainable ML Interface QuestionsFrom formative research
What if I added…
What are similar homes? ML
Why does this house cost that much?
What is the difference between these two?
Where is it wrong?
Explainable ML Interface QuestionsFrom formative research
What if I added…
What are similar homes? ML
Why does this house cost that much?
What is the difference between these two?
Where is it wrong?
Explainable ML Interface QuestionsFrom formative research
Why does this house cost that much?
What is most important?
What is the difference between these two?
What if I added…
What are similar homes?
Where is it wrong?ML
Explainable ML Interface QuestionsFrom formative research
Why does this house cost that much?
What is most important?
What is the difference between these two?
What if I added…
What are similar homes?
Where is it wrong?ML
Explainable ML Interface QuestionsFrom formative research
Why does this house cost that much?
What is most important?
What is the difference between these two?
What if I added…
What are similar homes?
Where is it wrong?
Explainable ML Interface QuestionsFrom formative research
Explainable ML InterfaceWhy does this house cost that much?
What is most important?
What is the difference between these two?
What if I added…
What are similar homes?
Where is it wrong?
Capabilities
Local instance explanations
Feature importance
Instance explanation comparisons
Counterfactuals
Nearest neighbors
Regions of error
C1
C2
C3
C4
C5
C6
From formative research
Explainable ML InterfaceWhy does this house cost that much?
What is most important?
What is the difference between these two?
What if I added…
What are similar homes?
Where is it wrong?
Capabilities
Local instance explanations
Feature importance
Instance explanation comparisons
Counterfactuals
Nearest neighbors
Regions of error
C1
C2
C3
C4
C5
C6
From formative research
Definitions + examples in the paper!
How to test our capabilities?
!30
Contribution 2: Design Probe
How to test our capabilities?
!30
Contribution 2: Design Probe
Goal: understand emerging practice of model interpretability
Design probe: “instrument that is deployed to find out about the unknown—returning with useful or interesting data.” Balance of design, social science, engineering
How to test our capabilities?
[Hutchinson, 2003]
!30
Contribution 2: Design Probe
Goal: understand emerging practice of model interpretability
!31
How does our design probe support our capabilities?