How many dimensions are required to find an adversarial example?. (arXiv:2303.14173v1 [cs.LG])

Past work exploring adversarial vulnerability have focused on situations
where an adversary can perturb all dimensions of model input. On the other
hand, a range of recent works consider the case where either (i) an adversary
can perturb a limited number of input parameters or (ii) a subset of modalities
in a multimodal problem. In both of these cases, adversarial examples are
effectively constrained to a subspace $V$ in the ambient input space
$\mathcal{X}$. Motivated by this, in this work …

adversarial adversary ambient case cases effectively find input problem space vulnerability work