Cross-lingual AMR Parsing on AMR German (DE) human-translated 2.0 (test)
0.746SmatchWLT+Structured-BART+MBSE-A silver
Evaluation Results
| Method | Links | |
|---|---|---|
| WLT+Structured-BART+MBSE-A silverLM=BART, Variant=average-Smatch, ensemble decoding2021.12 | 0.746 | |
| WLT+Structured-BART+MBSE-G silverLM=BART, Variant=greedy-selection2021.12 | 0.739 | |
| Ensemble-5 distillation + MBSE-A silver + ens. dec.LM=MBART, Decoding=ensemble decoding2021.12 | 0.737 | |
| Ensemble-5 distillation + MBSE-A silverLM=MBART, Distillation=APT + Structured-BART + SPRING1 + SPRING2 + AMRBART2021.12 | 0.735 | |
| Ensemble-4 distillation + MBSE-G silver + AMR2Text + ens. dec.LM=MBART, Decoding=ensemble decoding2021.12 | 0.732 | |
| Cai et al. (2021a)LM=mBARTmmt, Training Data=320K silver AMR2021.12 | 0.731 | |
| Ensemble-4 distillation + MBSE-G silver + AMR2TextLM=MBART, Augmentation=AMR2Text2021.12 | 0.729 | |
| Ensemble-4 distillation + MBSE-G silverLM=mBART, Distillation=APT + Structured-BART + SPRING1 + SPRING22021.12 | 0.725 | |
| Xu et al. (2021)2021.12 | 0.705 | |
| Structured-mBART-baselineLM=MBART2021.12 | 0.699 | |
| Procopio et al. (2021)LM=mBARTmt, Training Data=5M parallel corpus2021.12 | 0.698 | |
| Uhrig et al. (2021)2021.12 | 0.676 | |
| Cai et al. (2021b)2021.12 | 0.64 | |
| Sheth et al. (2021)LM=XLMR, Training Data=85K silver AMR2021.12 | 0.627 | |
| Blloshmi et al. (2020)2021.12 | 0.53 |