History of Αντίμετρα

Last update 2020-06-04 15:26:20

Creation date 2020-06-04 00:16:04

• 
  xr on 2020-06-04 15:26:20Informe

  Evasion attacks

  Multiple classifier systems, Learning with Invariances (SVMs), Game Theory (SVMs)

  Poisoning or causative or pollution attacks

  Magnet,Data sanitization, Robust learning (PCA), Randomization, information hiding, security by obscurity

• 
  xr on 2020-06-04 00:27:37Informe

  Poisoning or causative or pollution attacks

  Magnet

  Backdoor or Trojan attacks

  Fine pruning, Neural Cleanse, Activation Clustering, Input Image Blurring, Multi-layer Tuning in Transfer Learning, Input anomaly detection, Retraining DNN, Preprocessing, Strip (Strong intentional perturbation)

• 
  xr on 2020-06-04 00:16:04Informe

  Mode

  70

  Nombre

  Αντίμετρα

  Viewer

  unlisted

  Editor

  xr-4qknqqmd

  Inference or privacy attacks

  Min-max game, Dropout, Model stalking, Differential privacy, Memguard

  Gradient-based attacks

  Grandient masking μηχανισμοί, Non-obfuscated gradients:Adversarial Training,Cascade Adversarial Training, Shatered gradients: Thermometer Encoding, Input Transformations, Local Intrinsic Dimensionality, Stochastic gradients: Stochastic Activation Pruning, Vanishing gradients: Pixel Defend, Local Gradients Smoothing (LGS)