Document Classification on 20 Newsgroups (test)
96.93AccuracyFRAGE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FRAGEMethod variant=with FRAGE2018.09 | 96.93 | — | |
| BaselineMethod variant=Original2018.09 | 96.49 | — | |
| BERT_BASE Fine-tuneBackbone=BERT_BASE, Evaluation Protocol=Fine-tuning, Total number of parameters=17x, Trained parameters per task=100%2019.02 | 92.8 | — | |
| BERT_BASE Variable FTBackbone=BERT_BASE, Evaluation Protocol=Variable Fine-tuning, Total number of parameters=9.9x, Trained parameters per task=52.9%2019.02 | 92.8 | — | |
| BERT_BASE AdaptersBackbone=BERT_BASE, Evaluation Protocol=Adapter-tuning, Total number of parameters=1.19x, Trained parameters per task=1.14%2019.02 | 91.7 | — | |
| No BERT baseline2019.02 | 91.1 | — | |
| Text GCN2018.09 | 86.34 | — | |
| SPAM (Linear, Order 3)Interaction Order=3, Model Architecture=Linear2022.05 | 85.2 | — | |
| SWEM2018.09 | 85.16 | — | |
| DNNModel Architecture=Neural2022.05 | 84.94 | — | |
| SPAM (Linear, Order 2)Interaction Order=2, Model Architecture=Linear2022.05 | 84.72 | — | |
| TF-IDF + LR2018.09 | 83.19 | — | |
| Linear (Order 1)Interaction Order=1, Model Architecture=Linear2022.05 | 82.38 | — | |
| Multinomial NBimplementation=Scikit-learn, parameters=default2020.10 | 82.28 | — | |
| CNNInitialization=pre-trained GloVe2018.09 | 82.15 | — | |
| LEAM2018.09 | 81.91 | — | |
| (D(.||.), ||.||F)-SSNMFrank=13, max iterations=502020.10 | 81.88 | 0.44 | |
| (D(.||.), D(.||.))-SSNMFrank=13, max iterations=502020.10 | 81.5 | 0.47 | |
| Graph-CNN-C2018.09 | 81.42 | — | |
| SVMimplementation=Scikit-learn, parameters=default2020.10 | 80.7 | 0.27 | |
| fastTextN-grams=bigrams2018.09 | 79.67 | — | |
| (||.||F, D(.||.))-SSNMFrank=13, max iterations=502020.10 | 79.51 | 0.38 | |
| fastTextN-grams=unigrams2018.09 | 79.38 | — | |
| (||.||F, ||.||F)-SSNMFrank=13, max iterations=502020.10 | 79.37 | 0.47 | |
| CNNInitialization=random2018.09 | 76.93 | — | |
| XGBoostModel Architecture=Trees2022.05 | 76.77 | — | |
| PTE2018.09 | 76.74 | — | |
| HyperGCLTraining label rate=1%, Augmentation strategy=A62022.10 | 75.52 | — | |
| LSTMInitialization=pre-trained GloVe2018.09 | 75.43 | — | |
| HyperGCLTraining label rate=1%, Augmentation strategy=A52022.10 | 74.81 | — | |
| HyperGCLTraining label rate=1%, Augmentation strategy=A22022.10 | 74.72 | — | |
| HyperGCLTraining label rate=1%, Augmentation strategy=A12022.10 | 74.68 | — | |
| HyperGCLTraining label rate=1%, Augmentation strategy=A42022.10 | 74.67 | — | |
| HyperGCLTraining label rate=1%, Augmentation strategy=A32022.10 | 74.63 | — | |
| HyperGCLTraining label rate=1%, Augmentation strategy=A02022.10 | 74.43 | — | |
| SetGNN (Con)Training label rate=1%, Variant=Contrastive2022.10 | 74.39 | — | |
| PV-DBOW2018.09 | 74.36 | — | |
| SetGNN (Self)Training label rate=1%, Variant=Self-supervised2022.10 | 73.91 | — | |
| SetGNNTraining label rate=1%2022.10 | 73.83 | — | |
| Bi-LSTM2018.09 | 73.18 | — | |
| ||.||F-NMF + SVMrank=13, max iterations=400, initialization=random2020.10 | 70.99 | 2.71 | |
| LSTMInitialization=random2018.09 | 65.71 | — | |
| PV-DM2018.09 | 51.14 | — |