thankyouforpointingthis...
51425b752a0b402ed3effc83fc4bbb74-AuthorFeedback.pdf
Other optimization problems for which the method could be24 demonstrated?] We did additional experiments with features extracted from a pre-trained neural network, andex25 matchedexactlythetesterrorofx ( 10%). Thankyouforpointingthisout.39 [Intuition for why the proposed approach works.]
cfee398643cbc3dc5eefc89334cacdc1-AuthorFeedback.pdf
We thank the reviewers for their insightful comments. A number of additional experiments were suggested. AQ produces models specifically optimized for robust25 few-shotadaptation. In Section 4.2, we note that43 perturbing support data optimizes the network for adversarial fine-tuning.