method
active
method:noise-plus-one-step-gradient-l2-attackNoise-plus-one-step gradient L2 attack
Primary attack used for toy model robustness.
Neighborhood — ranked by edge-count
Papers (1)
paper
- Adversarial Examples Are Not Bugs, They Are Superpositionintroducesmentions
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Used instead of gradient descent attacks due to gradient masking.
- Used for updating hidden state expectations; provides dynamical process theory testable against neuronal data
- Frequent-task updates overwrite rare-task features.
- Optimization procedure for simultaneously updating action selection and perception; uses step size ζ (default 4).
- Gradient balancing by solving multi-objective optimization for minimum-norm aggregated gradient.
- Unsatisfying reduction.
- Reduces superposition but needs 80% L2 norm.
- Theory that cellular electrochemical gradients are the evolutionary root of sentience; a Level 4 substrate constraint.