Imperio: Language-Guided Backdoor Attacks for Arbitrary Model Control.

Ka-Ho Chow, Wenqi Wei, Lei Yu

Browse the full IJCAI paper archive.