Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Multi-task Learning research benchmarks – Page 2 · SOTA2 Research

Multi-task Learning

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Progressive learning

1.791PMTL Score

8

Jun 16, 2026

90.9Normalized Performance

8

Apr 15, 2026

48.2Worst Performance

7

Jun 11, 2026

0.129L1 Loss

7

Apr 3, 2026

0.315L1 Loss

7

Apr 3, 2026

0Am

7

Feb 26, 2026

Taskonomy tiny (test)

89.75Object Classification

7

Feb 26, 2026

PASCAL-Context 5 tasks

81.56HV Score

7

Feb 26, 2026

CIFAR-100 (test)

0.01HV

7

Feb 26, 2026

3.7ΔT1

7

Feb 27, 2026

Mixture of 6 task-level SFT tasks (test)

0.508Test Loss

6

Jun 4, 2026

Office-31 (test)

89.06Accuracy (Domain A)

6

Feb 26, 2026

46.33SemSeg Acc

5

Feb 26, 2026

378.286HV

4

Jun 4, 2026

200Epochs

4

Feb 26, 2026

Multi-MNIST Fashion F+M

100Epochs

4

Feb 26, 2026

Multi-Task Learning

158.17Initialization Time

3

Jun 19, 2026

VKITTI2 -> Cityscapes 8 classes

-13.37Delta (ASD)

3

Feb 26, 2026

Synthia -> Cityscapes 16 classes

-30.49Delta (ASD)

3

Feb 26, 2026

91Speed (fps)

3

Feb 26, 2026

Page 2 of 2

Popular datasets

NYU Cityscapes Pascal Context NYUD Taskonomy

Follow for updates

Join SOTA2 Contact