Unverified50% confidenceFactExact time
The PGD adversarial training proposed by Madry et al. is a classic adversarial training implementation; it requires additional computation of adversarial examples at each step, making training far more costly than standard training.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAdversarial training (AT) improves model robustness by generating adversarial examples during training and including them in training, but it incurs enormous memory overhead and involves a trade-off between clean accuracy and robust accuracy.71% similarUnverifiedAdversarial training has the limitations of poor cross-attack-type generalization and high training cost.68% similarUnverifiedResistance training prescriptions generally require comprehensive consideration of key variables such as training intensity, training volume, training frequency, exercise selection, and inter-set rest periods.66% similarUnverifiedGAIL借鉴GAN思想，用对抗训练隐式恢复专家行为分布，在连续控制任务上取得接近IRL的效果且计算更高效66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1114007API
curl https://kongchang.com/api/v1/knowledge/claims/1114007MCP
get_claim(id=1114007)