Multi-task Classification research benchmarks

Benchmarks

Dataset NameSOTA methodMetricTrendResultsLast Updated

30 image classification tasks average (test)

Fine-tuned

93.1Average Accuracy

Jun 23, 2026

ImageCLEF

STL

95Ao

Feb 26, 2026

Office-Caltech

STL

98.61Accuracy (Task o)

Feb 26, 2026

Office-Home

STL

88.25Metric Ao

Feb 26, 2026

GLUE MRPC, RTE, CoLA (test, val)

DSGD

88.95MRPC Accuracy

May 1, 2026

Landmine (test)

RMB-CLE-via-LGBM

63.5F1 Score

Feb 26, 2026

Bank (test)

RMB-CLE-via-LGBM

73.8F1 Score

Feb 26, 2026

Adult-Race (test)

RMB-CLE-via-LGBM

0.815F1 Score

Feb 26, 2026

Adult-Gender (test)

RMB-CLE-via-LGBM

81.4F1 Score

Feb 26, 2026

CelebA (test)

Double-Clip MGDA

88.41Averaged Accuracy

May 8, 2026

Skin-Lesion (gamma = 0%)

Association Graph Learning

85.98Overall Accuracy

Feb 26, 2026

Skin-Lesion gamma = 33%

Association Graph Learning

16.58Classification Component Am

Feb 26, 2026

Skin-Lesion gamma = 67%

Association Graph Learning

10.82Metric Am

Feb 26, 2026