Model Card
Model Cards are a type of governance artifact that document basic technical information about an ML model, which is relevant regardless of where and how the model is going to be deployed; Model Cards can be created by model builders before a final end application for the model has been identified.
Model Cards are designed to give stakeholders visibility into how a model works and what it does, along with potential technical limitations and risks.
Dataset Cards are a similar type of technical artifact we are pioneering for datasets (in the same vein of Datasheets and Nutrition Labels for Datasets), designed to provide some basic information about what a dataset is and contains, its source, and its potential limitations for use in training (or testing) an ML system.
Both Model Cards and Dataset Cards are modular components of our proposed AI Impact Assessment, in addition to being standalone artifacts that could be useful on their own.
1
A
2
B
3
C
4
D
5
E
6
F
7
G
8
H
9
I
12
L
13
M
16
P
18
R
19
S
20