Matches in SemOpenAlex for { <https://semopenalex.org/work/W2940648411> ?p ?o ?g. }
- W2940648411 abstract "Adversarial training, in which a network is trained on adversarial examples, is one of the few defenses against adversarial attacks that withstands strong attacks. Unfortunately, the high cost of generating strong adversarial examples makes standard adversarial training impractical on large-scale problems like ImageNet. We present an algorithm that eliminates the overhead cost of generating adversarial examples by recycling the gradient information computed when updating model parameters. Our free adversarial training algorithm achieves comparable robustness to PGD adversarial training on the CIFAR-10 and CIFAR-100 datasets at negligible additional cost compared to natural training, and can be 7 to 30 times faster than other strong adversarial training methods. Using a single workstation with 4 P100 GPUs and 2 days of runtime, we can train a robust model for the large-scale ImageNet classification task that maintains 40% accuracy against PGD attacks. The code is available at this https URL." @default.
- W2940648411 created "2019-05-03" @default.
- W2940648411 creator A5001523654 @default.
- W2940648411 creator A5009839197 @default.
- W2940648411 creator A5021900923 @default.
- W2940648411 creator A5037237895 @default.
- W2940648411 creator A5043788981 @default.
- W2940648411 creator A5055264952 @default.
- W2940648411 creator A5060687985 @default.
- W2940648411 creator A5071538560 @default.
- W2940648411 creator A5083617223 @default.
- W2940648411 date "2019-04-29" @default.
- W2940648411 modified "2023-09-25" @default.
- W2940648411 title "Adversarial Training for Free" @default.
- W2940648411 cites W1673923490 @default.
- W2940648411 cites W2108598243 @default.
- W2940648411 cites W2117539524 @default.
- W2940648411 cites W2401231614 @default.
- W2940648411 cites W2460937040 @default.
- W2940648411 cites W2604147826 @default.
- W2940648411 cites W2618043096 @default.
- W2940648411 cites W2640329709 @default.
- W2940648411 cites W2765233338 @default.
- W2940648411 cites W2765384636 @default.
- W2940648411 cites W2766462876 @default.
- W2940648411 cites W2774018344 @default.
- W2940648411 cites W2778624544 @default.
- W2940648411 cites W2783555701 @default.
- W2940648411 cites W2786104118 @default.
- W2940648411 cites W2786118190 @default.
- W2940648411 cites W2787708942 @default.
- W2940648411 cites W2787733970 @default.
- W2940648411 cites W2788232741 @default.
- W2940648411 cites W2791953061 @default.
- W2940648411 cites W2884821828 @default.
- W2940648411 cites W2892354372 @default.
- W2940648411 cites W2893001471 @default.
- W2940648411 cites W2897355816 @default.
- W2940648411 cites W2898152545 @default.
- W2940648411 cites W2899692219 @default.
- W2940648411 cites W2908392948 @default.
- W2940648411 cites W2911634294 @default.
- W2940648411 cites W2942836458 @default.
- W2940648411 cites W2962821226 @default.
- W2940648411 cites W2962872506 @default.
- W2940648411 cites W2962933129 @default.
- W2940648411 cites W2962943487 @default.
- W2940648411 cites W2963158386 @default.
- W2940648411 cites W2963207607 @default.
- W2940648411 cites W2963249138 @default.
- W2940648411 cites W2963389226 @default.
- W2940648411 cites W2963540169 @default.
- W2940648411 cites W2963626025 @default.
- W2940648411 cites W2963744840 @default.
- W2940648411 cites W2963855547 @default.
- W2940648411 cites W2963857521 @default.
- W2940648411 cites W2963920068 @default.
- W2940648411 cites W2963952467 @default.
- W2940648411 cites W2964014389 @default.
- W2940648411 cites W2997502936 @default.
- W2940648411 cites W9657784 @default.
- W2940648411 hasPublicationYear "2019" @default.
- W2940648411 type Work @default.
- W2940648411 sameAs 2940648411 @default.
- W2940648411 citedByCount "0" @default.
- W2940648411 crossrefType "posted-content" @default.
- W2940648411 hasAuthorship W2940648411A5001523654 @default.
- W2940648411 hasAuthorship W2940648411A5009839197 @default.
- W2940648411 hasAuthorship W2940648411A5021900923 @default.
- W2940648411 hasAuthorship W2940648411A5037237895 @default.
- W2940648411 hasAuthorship W2940648411A5043788981 @default.
- W2940648411 hasAuthorship W2940648411A5055264952 @default.
- W2940648411 hasAuthorship W2940648411A5060687985 @default.
- W2940648411 hasAuthorship W2940648411A5071538560 @default.
- W2940648411 hasAuthorship W2940648411A5083617223 @default.
- W2940648411 hasConcept C104317684 @default.
- W2940648411 hasConcept C111919701 @default.
- W2940648411 hasConcept C113775141 @default.
- W2940648411 hasConcept C119857082 @default.
- W2940648411 hasConcept C121332964 @default.
- W2940648411 hasConcept C153294291 @default.
- W2940648411 hasConcept C154945302 @default.
- W2940648411 hasConcept C177264268 @default.
- W2940648411 hasConcept C185592680 @default.
- W2940648411 hasConcept C199360897 @default.
- W2940648411 hasConcept C2776760102 @default.
- W2940648411 hasConcept C2777211547 @default.
- W2940648411 hasConcept C2778755073 @default.
- W2940648411 hasConcept C2779960059 @default.
- W2940648411 hasConcept C37736160 @default.
- W2940648411 hasConcept C41008148 @default.
- W2940648411 hasConcept C55493867 @default.
- W2940648411 hasConcept C62520636 @default.
- W2940648411 hasConcept C63479239 @default.
- W2940648411 hasConcept C67953723 @default.
- W2940648411 hasConceptScore W2940648411C104317684 @default.
- W2940648411 hasConceptScore W2940648411C111919701 @default.
- W2940648411 hasConceptScore W2940648411C113775141 @default.
- W2940648411 hasConceptScore W2940648411C119857082 @default.
- W2940648411 hasConceptScore W2940648411C121332964 @default.