Livoa
Discord
Pricing
English
Sign In
Black-box model
x
→
M
φ
→
y
x
= (
x
1
, ...,
x
n
)
x
i
: input instance
M
φ
Visualization
M
φ
"What happens with the prediction
y
i
if we change slightly the features of
x
i
?"
Local explanations
M
φ
"Feature
x
2
has a 90% importance in
y'
"
Feature relevance
M
φ
"Explanatory examples for the model:"
○ -
x
A
→
y
A
○ -
x
B
→
y
B
○ -
x
C
→
y
C
Explanations by example
M
φ
"The output for
x
i
is
y
i
because
x
3
>
γ'
"
Text explanations
M
φ
F
→
G
Model simplification
wdefg
by Man T
Use this design
0
0 uses
