Let x

be the feature vector $($x$1,$x$2,$x$3,$x$4)$

$,$ with its true label y

$.$ Let x$$

be x

$,$ with the $4$th feature missing, i$.$e$.,$ x$=($x$1,$x$2,$x$3,)$

$,$ where $\text{'}*\text{'}$ means a missing feature. What best describes attribute inference?

Question $3$Answer

a$.$

Querying the model to determine if x

was part of the training data

b$.$

Querying the model to find the missing feature in x$$