Mathematical Information Retrieval on CLEF ARQMath 2
32.9NDCG'ColBERT
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ColBERTTraining stage=fully-trained2022.03 | 32.9 | 12.8 | 21.3 | 13.6 | 69.5 | |
| DPRTraining stage=fully-trained2022.03 | 27 | 8.7 | 15.2 | 9.7 | 66.3 | |
| MIRMU CompuBERT2022.03 | 26.2 | 8.3 | 13.5 | 8.7 | 69.2 | |
| TU_DBS ColARQBERT2022.03 | 22.5 | 7.3 | 13.1 | — | — | |
| further-pretrained BERTPre-trained backbone=BERT, Fine-tuning epochs=1, Further-pretraining epochs=72022.03 | 20 | 6 | 13 | 8.1 | 46.8 | |
| further-pretrained BERTPre-trained backbone=BERT, Fine-tuning epochs=1, Further-pretraining epochs=32022.03 | 19.5 | 5.8 | 12.6 | 7.3 | 45 | |
| TU_DBS ColBERT2022.03 | 18.3 | 5.3 | 11 | — | — | |
| SciBERTPre-trained backbone=SciBERT, Fine-tuning epochs=12022.03 | 17.6 | 5.6 | 13.4 | 7.3 | 42.3 | |
| BERTPre-trained backbone=BERT, Fine-tuning epochs=12022.03 | 17.4 | 5.1 | 11.6 | 7.3 | 45.3 | |
| FormulaEmb2022.03 | 16.1 | 5.9 | 19.7 | — | — | |
| TU_DBS ColSciBERT2022.03 | 2.8 | 0.4 | 0.9 | 0.9 | 17.5 |