AI Paper Atlas

Research map / Learning theory and algorithms / Combinatorial and search optimization

LDC-MTL: Balancing Multi-Task Learning through Scalable Loss Discrepancy Control

Peiyao Xiao, Chaosheng Dong, Shaofeng Zou, Kaiyi Ji

ECCV 2026 · Learning theory and algorithms › Combinatorial and search optimization · approach: Multi-objective optimization · 9 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace linear scalarization with bilevel loss discrepancy controlLDC-MTL · CelebA (40-task)Δm%: 4.15 -> -1.31Improved
replace linear scalarization with bilevel loss discrepancy controlLDC-MTL · QM9 (11-task)Δm%: 177.6 -> 49.5Improved
use LDC-MTL instead of the strongest baseline FairGradLDC-MTL · Cityscapes (2-task)Δm%: 5.18 -> -0.57Improved
use LDC-MTL instead of GO4AlignLDC-MTL · NYU-v2 (3-task)Δm%: -6.08 -> -4.40Hurt
loss discrepancy control reduces spread of task lossesLDC-MTL vs LS · CelebA (40-task)mean task loss: 0.287 -> 0.189 (std 0.195 -> 0.133)Improved
add dynamic loss discrepancy control instead of weight-swept LSLDC-MTL · Cityscapes (2-task)Δm%: 5.13 -> -0.57Improved
increase penalty constant lambda from 0.02 to 0.1 (ablation)LDC-MTL · Cityscapes (2-task)Δm%: 1.96 -> -0.57Improved
vary lambda among 0.005, 0.008, 0.01, 0.02 (ablation)LDC-MTL · CelebA (40-task)Δm%: -1.27, -1.16, -1.31, -0.96No clear effect
randomly reorder task losses before computing the upper-level gap (ablation)LDC-MTL · CelebA (40-task)Δm%: -1.31 -> -1.30, -1.00, -1.62No clear effect
use single-loop first-order update (drop the inner loop) for O(1) costLDC-MTL · training timetime vs LS: at most 1.11x increase vs LSNo clear effect
use O(1) LDC-MTL instead of O(K) gradient manipulationLDC-MTL vs Nash-MTL · CelebA (40-task)training time: approximately 12x less training timeImproved
use dynamic task weights whose final values are near equal rather than fixed LSLDC-MTL · QM9 (11-task)Δm%: 177.6 -> 49.5Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy