Exponential Moving Average (EMA)

Used to estimate gradients dynamically, with forgetting rate β.

Neighborhood — ranked by edge-count

method

DB-MTL (Dual-Balancing Multi-Task Learning)
implements
The proposed method combining loss-scale balancing via logarithm transformation and gradient-magnitude balancing via maximum-norm normalization.

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Exponential Moving Averagemethod0.859
Used in DB-MTL to estimate batch gradient expectations dynamically
Exponential Moving Average Target Networkmethod0.799
Used in self-prior training as a slow target network for KL regularization
Earth Mover's Distance (EMD)method0.747
Primary quantitative measure of distributional divergence between natural and intervened representations
Microelectrode array (MEA)method0.705
Device to record and stimulate electrical activity of neural cultures.
EI and normalized EI could serve as a unified metric for out-of-distribution generalization.claim0.699
Conjecture that maximizing EI yields causal representations invariant to distribution shifts.
DB-MTL with EMA forgetting rate β in a wide range performs better than without EMA (β=0) on Office-31.finding0.694
Effect of EMA forgetting rate on performance.
Equal Weighting (EW)method0.692
Baseline that minimizes sum of task losses with equal weights.
For small CL loss weights epsilon, IIA is maintained (potentially improved) while EMD decreases in Boundless DAS on a 7B LLMfinding0.689
Empirical result showing the CL loss can reduce divergence without sacrificing interpretability accuracy