method
active
method:exponential-moving-averageExponential Moving Average
Used in DB-MTL to estimate batch gradient expectations dynamically
Neighborhood — ranked by edge-count
Methods (1)
method
- maximum-norm gradient normalizationimplementsTraining-free technique normalizing all task gradients to the maximum gradient norm magnitude
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 to estimate gradients dynamically, with forgetting rate β.
- Used in self-prior training as a slow target network for KL regularization
- Measure of how far cells travel during development; increased by stress sharing.
- Averages AUROC over 500 simulations drawing 1,000 precomputed rollouts per α.
- Paper's main proposal: SAE trained on the mean activation across a turn.
- Metric estimating basin-boundary fractal dimension from how the fraction of differing-outcome pixel pairs scales with separation.
- Attribution graphs with nodes per turn per layer.