# Multimodal Classification on CelebA

Source: [Multimodal Classification on CelebA](<https://www.sota2.com/research/sota/multimodal-classification-on-celeba>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-03T07:44:15.136Z

Evaluation of model performance on the CelebA dataset using Multimodal Classification, reporting accuracy across image, tabular, and combined modalities.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Multimodal Classification |
| Dataset (including any stated split) | CelebA |
| Metric | Accuracy (Image) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 19 of 19 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Accuracy (Image)](<https://www.sota2.com/research/sota/multimodal-classification-on-celeba.md?metric=Accuracy%20(Image)>): Higher is better; unit: PERCENT.

- [Accuracy (Multi)](<https://www.sota2.com/research/sota/multimodal-classification-on-celeba.md?metric=Accuracy%20(Multi)>): Higher is better; unit: PERCENT.

- [Accuracy (Tabular)](<https://www.sota2.com/research/sota/multimodal-classification-on-celeba.md?metric=Accuracy%20(Tabular)>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy (Image) | Result date | Report paper |
| --- | --- | --- | --- |
| GAAL\* | 83.13 | Not recorded | 69cf49780636e5bf4afcd987 |
| GAAL | 82.22 | Not recorded | 69cf49780636e5bf4afcd987 |
| CHARMS | 82.2 | Not recorded | 69cf49780636e5bf4afcd987 |
| Resnet50 | 81.46 | Not recorded | 69cf49780636e5bf4afcd987 |
| ReconBoost | 81.33 | Not recorded | 69cf49780636e5bf4afcd987 |
| DI-MML | 81.3 | Not recorded | 69cf49780636e5bf4afcd987 |
| KD | 81.18 | Not recorded | 69cf49780636e5bf4afcd987 |
| LFM | 81.1 | Not recorded | 69cf49780636e5bf4afcd987 |
| MMCL | 80.41 | Not recorded | 69cf49780636e5bf4afcd987 |
| FMR | 80.03 | Not recorded | 69cf49780636e5bf4afcd987 |
| MMPareto | 79.41 | Not recorded | 69cf49780636e5bf4afcd987 |
| MLA | 79.15 | Not recorded | 69cf49780636e5bf4afcd987 |
| MFH | 75.07 | Not recorded | 69cf49780636e5bf4afcd987 |
| CF | 56.54 | Not recorded | 69cf49780636e5bf4afcd987 |
| OGM | 53.28 | Not recorded | 69cf49780636e5bf4afcd987 |

### GAAL\*

Accuracy (Image): 83.13. Result ID: 69cf6f0447f34ce924b8fb60.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Test-time Tabular Missing","Pretrained weight setting":"Same as test-time tabular missing methods"}. This record does not assert independent reproduction.

### GAAL

Accuracy (Image): 82.22. Result ID: 69cf6f0447f34ce924b8fb5a.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Proposed Method"}. This record does not assert independent reproduction.

### CHARMS

Accuracy (Image): 82.2. Result ID: 69cf6f0447f34ce924b8fb5f.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Test-time Tabular Missing"}. This record does not assert independent reproduction.

### Resnet50

Accuracy (Image): 81.46. Result ID: 69cf6f0447f34ce924b8fb4e.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62321e209131ce9f8328.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Unimodal Learning"}. This record does not assert independent reproduction.

### ReconBoost

Accuracy (Image): 81.33. Result ID: 69cf6f0447f34ce924b8fb58.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73371e209131cea5b3a6.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Multimodal Gradient Conflict"}. This record does not assert independent reproduction.

### DI-MML

Accuracy (Image): 81.3. Result ID: 69cf6f0447f34ce924b8fb57.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74541e209131cea611ec.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Multimodal Gradient Conflict"}. This record does not assert independent reproduction.

### KD

Accuracy (Image): 81.18. Result ID: 69cf6f0447f34ce924b8fb5b.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa61b81e209131ce9f5b87.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Test-time Tabular Missing"}. This record does not assert independent reproduction.

### LFM

Accuracy (Image): 81.1. Result ID: 69cf6f0447f34ce924b8fb59.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Multimodal Gradient Conflict"}. This record does not assert independent reproduction.

### MMCL

Accuracy (Image): 80.41. Result ID: 69cf6f0447f34ce924b8fb5e.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ee41e209131cea42afe.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Test-time Tabular Missing"}. This record does not assert independent reproduction.

### FMR

Accuracy (Image): 80.03. Result ID: 69cf6f0447f34ce924b8fb5d.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Test-time Tabular Missing"}. This record does not assert independent reproduction.

### MMPareto

Accuracy (Image): 79.41. Result ID: 69cf6f0447f34ce924b8fb55.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73501e209131cea5bbff.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Multimodal Gradient Conflict"}. This record does not assert independent reproduction.

### MLA

Accuracy (Image): 79.15. Result ID: 69cf6f0447f34ce924b8fb56.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72bd1e209131cea5848f.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Multimodal Gradient Conflict"}. This record does not assert independent reproduction.

### MFH

Accuracy (Image): 75.07. Result ID: 69cf6f0447f34ce924b8fb5c.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ea91e209131cea412e3.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Test-time Tabular Missing"}. This record does not assert independent reproduction.

### CF

Accuracy (Image): 56.54. Result ID: 69cf6f0447f34ce924b8fb50.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Tabular-image Fusion"}. This record does not assert independent reproduction.

### OGM

Accuracy (Image): 53.28. Result ID: 69cf6f0447f34ce924b8fb54.

[Reporting paper](<https://www.sota2.com/research/paper/69cf49780636e5bf4afcd987.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6c161e209131cea32100.md>)

Recorded evaluation configuration: {"Learning Paradigm":"Multimodal Gradient Conflict"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
