# Novel Class Discovery on CIFAR-100 (50/50 seen/novel split)

Source: [Novel Class Discovery on CIFAR-100 (50/50 seen/novel split)](<https://www.sota2.com/research/sota/novel-class-discovery-on-cifar-100-50-50-seen-novel-split>)

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

Data updated: 2026-02-26T05:28:27.302Z

Evaluation of Novel Class Discovery performance on the CIFAR-100 dataset using a specific split where 50% of classes are seen and 50% are novel, measured by Accuracy.

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 | Novel Class Discovery |
| Dataset (including any stated split) | CIFAR-100 (50/50 seen/novel split) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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](<https://www.sota2.com/research/sota/novel-class-discovery-on-cifar-100-50-50-seen-novel-split.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| TRSSL | 57.5 | Not recorded | 68fa6d201e209131cea37fac |
| UNO | 52.9 | Not recorded | 68fa6d201e209131cea37fac |
| RankStats+ | 44.1 | Not recorded | 68fa6d201e209131cea37fac |
| RankStats | 39.2 | Not recorded | 68fa6d201e209131cea37fac |
| DTC | 35.9 | Not recorded | 68fa6d201e209131cea37fac |
| k-means | 28.3 | Not recorded | 68fa6d201e209131cea37fac |

### TRSSL

Accuracy: 57.5. Result ID: 698f94223a6c314849fa4198.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d201e209131cea37fac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d201e209131cea37fac.md>)

[Implementation](<https://github.com/nayeemrizve/TRSSL>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### UNO

Accuracy: 52.9. Result ID: 698f94223a6c314849fa4197.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d201e209131cea37fac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ae91e209131cea2afa3.md>)

[Implementation](<https://ncd-uno.github.io/>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### RankStats+

Accuracy: 44.1. Result ID: 698f94223a6c314849fa4196.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d201e209131cea37fac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67091e209131cea14026.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### RankStats

Accuracy: 39.2. Result ID: 698f94223a6c314849fa4195.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d201e209131cea37fac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67091e209131cea14026.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### DTC

Accuracy: 35.9. Result ID: 698f94223a6c314849fa4194.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d201e209131cea37fac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66441e209131cea0f56d.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### k-means

Accuracy: 28.3. Result ID: 698f94223a6c314849fa4193.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6d201e209131cea37fac.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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