Abstract: Existing unlearning algorithms in text-to-image generative models often fail to preserve the knowledge of semantically related concepts when removing specific target concepts—a challenge ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results