Text Categorization on 20NEWS (test)
81.37AccuracyNoiseAL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| NoiseALBackbone=BERT, Noise Type=real-noisy2025.04 | 81.37 | — | |
| CLBackbone=BERT, Noise Type=real-noisy2025.04 | 80.83 | — | |
| ELRBackbone=BERT, Noise Type=real-noisy2025.04 | 80.71 | — | |
| DyGenBackbone=BERT, Noise Type=real-noisy2025.04 | 80.5 | — | |
| Co-TeachingBackbone=BERT, Noise Type=real-noisy2025.04 | 80.43 | — | |
| SCEBackbone=BERT, Noise Type=real-noisy2025.04 | 80.42 | — | |
| SelfMixBackbone=BERT, Noise Type=real-noisy2025.04 | 80.16 | — | |
| BERTBackbone=BERT, Noise Type=real-noisy2025.04 | 80.06 | — | |
| Multinomial Naive Bayes2016.06 | 68.51 | — | |
| GC32Architecture=GC32, Support (K)=5, Graph=16-NN graph2016.06 | 68.26 | — | |
| Softmax2016.06 | 66.28 | — | |
| Linear SVM2016.06 | 65.9 | — | |
| FC2500-FC500Architecture=FC2500-FC5002016.06 | 65.76 | — | |
| FC2500Architecture=FC25002016.06 | 64.64 | — | |
| ApproxRepSet2019.04 | — | 23.82 | |
| DeepSets2019.04 | — | 38.88 | |
| NN-attentionpooling=attention2019.04 | — | 28.73 | |
| NN-maxpooling=max2019.04 | — | 32.15 | |
| NN-meanpooling=mean2019.04 | — | 38.4 | |
| RepSet2019.04 | — | 22.98 | |
| S-WMD2019.04 | — | 26.8 | |
| Set-Transformer2019.04 | — | 30.01 | |
| WMD2019.04 | — | 26.8 |