Hierarchical Text Classification on RCV1 v2 (test)
87.31Micro-F1HILL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| HILLText Encoder=BERT, Learning Paradigm=Contrastive2024.03 | 87.31 | 70.12 | |
| UMEEvaluation Protocol=supervised2026.03 | 86.88 | 69.7 | |
| HiTINText Encoder=BERT2023.05 | 86.71 | 69.95 | |
| HGCLRText Encoder=BERT, Learning Paradigm=Contrastive2024.03 | 86.49 | 68.31 | |
| HGCLRText Encoder=BERT2023.05 | 86.49 | 68.31 | |
| BERT+HiMatchText Encoder=BERT, Learning Paradigm=Supervised2024.03 | 86.33 | 68.66 | |
| HiMatchText Encoder=BERT2023.05 | 86.33 | 68.66 | |
| BERT (Chen et al., 2021)Text Encoder=BERT, Learning Paradigm=Supervised2024.03 | 86.26 | 67.35 | |
| BERTText Encoder=BERT, Learning Paradigm=Supervised2024.03 | 85.65 | 67.02 | |
| BERTText Encoder=BERT2023.05 | 85.65 | 67.02 | |
| BERT+HiAGMText Encoder=BERT, Learning Paradigm=Supervised2024.03 | 85.58 | 67.93 | |
| HiAGMText Encoder=BERT2023.05 | 85.58 | 67.93 | |
| BERT+HTCInfoMaxText Encoder=BERT, Learning Paradigm=Supervised2024.03 | 85.53 | 67.09 | |
| HTCInfoMaxText Encoder=BERT2023.05 | 85.53 | 67.09 | |
| HiMatchText Encoder=TextRCNN, Learning Paradigm=Supervised2024.03 | 84.73 | 64.11 | |
| HiAGMText Encoder=TextRCNN, Learning Paradigm=Supervised2024.03 | 83.96 | 63.35 | |
| HTCInfoMaxText Encoder=TextRCNN, Learning Paradigm=Supervised2024.03 | 83.51 | 62.71 | |
| TextRCNNText Encoder=TextRCNN, Learning Paradigm=Supervised2024.03 | 81.57 | 59.25 | |
| GPT-4Evaluation Protocol=zero-shot2026.03 | 56.12 | 36.54 | |
| LlaMa-3Evaluation Protocol=zero-shot2026.03 | 54.67 | 28.5 |