finding

active

finding:db-mtl-achieves-p-1-15-0-16-on-nyuv2-outperforming-all-baselines-including-state-of-the-art

DB-MTL achieves ∆p = +1.15±0.16 on NYUv2, outperforming all baselines including state-of-the-art

Primary empirical validation on scene understanding task

Source paper

extracted_from

Dual-Balancing for Multi-Task Learning

(2023) · Baijiong Lin · Weisen Jiang · Feiyang Ye · Yu Zhang +5

Neighborhood — ranked by edge-count

Claims (2)

claim

DB-MTL is a simple yet effective method that addresses both loss-scale and gradient-magnitude imbalances.
supports
Core claim of the paper.
Task balancing requires simultaneous consideration of both loss scales and gradient magnitudes
supports
Core interpretive position of DB-MTL: complementarity of loss and gradient perspectives

Communities (3)

community

Dual-balancing multi-task learning
members_of
DB-MTL jointly balances loss scale and gradient magnitude, benchmarked on NYUv2 and Office-31.
Dual balancing multi-task learning
members_of
DB-MTL combines loss-scale and gradient-magnitude balancing, benchmarked across NYUv2, Cityscapes, QM9, and Office datasets.
Multi-task learning gradient balancing
members_of
Methods addressing loss-scale and gradient-magnitude imbalances in multi-task learning, with DB-MTL achieving state-of-the-art results on dense prediction benchmarks like NYUv2.

Related by similarity (8)

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.

DB-MTL with SegNet backbone achieves Δp = +8.91 on NYUv2, best among all methods.finding0.872
Performance with a different backbone network.
Using only loss-scale balancing (log transformation) yields Δp = +0.06±0.09 on NYUv2.finding0.823
Ablation study component effectiveness.
DB-MTL has similar per-epoch running time to gradient balancing methods on NYUv2, slower than loss balancing methods.finding0.810
Computational efficiency comparison.
Combining loss-scale and gradient-magnitude balancing achieves Δp = +1.15±0.16 on NYUv2.finding0.783
Full DB-MTL ablation result.
DB-MTL training losses decrease smoothly and gradient norms are lower than EW on NYUv2, indicating training stability.finding0.769
Training stability analysis.
DB-MTL with EMA forgetting rate β in a wide range performs better than without EMA (β=0) on Office-31.finding0.764
Effect of EMA forgetting rate on performance.
On Qwen3-1.7B, MDS achieves ϕ1,C,↑ = 5.0 (SJTs) vs P2 at 4.7, and ϕ1,C,↓ = 1.4 (SJTs) vs P2 at 3.6finding0.761
Specific consciousness sweep result for Qwen3-1.7B from Table 6 demonstrating strong bidirectional steering
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.quote0.757
Concise summary of the DB-MTL method from the abstract.