Loss-scale balancing and gradient-magnitude balancing are complementary and combining them achieves the best performance.

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

• Task balancing requires simultaneous consideration of both loss scales and gradient magnitudesclaim0.881
  Core interpretive position of DB-MTL: complementarity of loss and gradient perspectives
• The proposed gradient-magnitude balancing method consistently outperforms GradNorm, as it guarantees equal gradient magnitudes and considers update magnitude.claim0.848
  Advantage over GradNorm.
• DB-MTL achieves loss-scale balancing by performing logarithm transformation on each task loss, and rescales gradient magnitudes by normalizing all task gradients to comparable magnitudes using the maximum gradient norm.quote0.838
  Concise summary of the DB-MTL method from the abstract.
• gradient-magnitude balancingconcept0.828
  Addressing disparity in gradient magnitudes across tasks at the gradient level
• The gradient-magnitude balancing method outperforms GradNorm on NYUv2, Cityscapes, Office-31, Office-Home.finding0.808
  Comparison of gradient-magnitude balancing with GradNorm.
• loss-scale balancingconcept0.804
  Addressing disparity in loss magnitudes across tasks at the loss level
• Setting aggregated gradient scaling factor to maximum gradient norm performs best for task balancingclaim0.796
  Empirical finding on choice of αk in gradient normalization strategy
• DB-MTL is a simple yet effective method that addresses both loss-scale and gradient-magnitude imbalances.claim0.788
  Core claim of the paper.