[New evaluation] <MALT attack>

[odeliamel]

Target model

• Model identifier(s):
  Wang2023Better_WRN-28-10
• Leaderboard(s):
  BenchmarkDataset.cifar_100, ThreatModel.Linf

New evaluation

• Link to adversarial images: <it should be a .pt tensor of the same size of the corresponding test set> :
  https://weizmannacil-my.sharepoint.com/:u:/g/personal/odelia_melamed_weizmann_ac_il/EVQ8UMa8S59Osu-v_9DEMGwByz6jqJjmrMDdVaVg83DrIg?e=leTkmD
• Robust accuracy::
  38.78%
• Author(s):
  Odelia Melamed, Gilad Yehudai, Adi Shamir
• Description of the method used: <it might be link to a paper, code or short description if the first two are not available>
  MALT Attack, accepted to Neurips24, fully described here: https://arxiv.org/pdf/2407.02240
  GitHub: https://github.com/odeliamel/MALT