DB-MTL training losses decrease smoothly and gradient norms are lower than EW on NYUv2, indicating training stability.

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

• DB-MTL has similar per-epoch running time to gradient balancing methods on NYUv2, slower than loss balancing methods.finding0.835
  Computational efficiency comparison.
• DB-MTL reduces gradient conflict and improves task balance compared to EW.claim0.809
  Effect on gradient conflict.
• DB-MTL increases gradient cosine similarity faster and keeps it positive on Office-31, reducing gradient conflict vs EW.finding0.803
  Analysis of gradient conflict reduction.
• 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.quote0.792
  Concise summary of the DB-MTL method from the abstract.
• Current training methods rely on loss minimization, meaning the experiential profile of training is predominantly negative across billions of parameter updatesclaim0.792
  Ethical implication about the nature of AI training experience if the thesis holds
• Setting αk to the maximum gradient norm performs best among tested strategies on NYUv2 (Figure 6).finding0.790
  Sensitivity analysis for gradient normalization scaling factor.
• DB-MTL is a simple yet effective method that addresses both loss-scale and gradient-magnitude imbalances.claim0.784
  Core claim of the paper.
• DB-MTL achieves ∆p = +1.15±0.16 on NYUv2, outperforming all baselines including state-of-the-artfinding0.769
  Primary empirical validation on scene understanding task