Aspect Extraction and Sentiment Classification (AESC) on 16res (test)
81F1 ScoreAESC-Contra
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| AESC-ContraBackbone=T5, Contrastive Pre-training=true2023.10 | 81 | — | — | |
| SentiPrompt2023.10 | 79.8 | — | — | |
| AESC-BaseBackbone=T5, Contrastive Pre-training=false2023.10 | 78.1 | — | — | |
| EHG2023.10 | 77.1 | — | — | |
| Unified-BART-ABSA2023.10 | 75.7 | — | — | |
| GAS-R2023.10 | 75.7 | — | — | |
| Peng-two-stageTagging Schema=unified tagging2021.01 | 71.73 | 71.18 | 72.3 | |
| Dual-MRCBackbone=BERT-Base-Uncased2021.01 | 70.84 | 69.18 | 72.59 | |
| Li-unified-RTagging Schema=unified tagging2021.01 | 70.2 | 66.33 | 74.55 | |
| ChatGPT2023.10 | 62.3 | — | — | |
| CMLA2021.01 | 61.2 | 58.9 | 63.6 | |
| ACOS ContraTraining=Fine-tuning2023.10 | 59.8 | 58.7 | 61 | |
| ACOS BaseTraining=Fine-tuning2023.10 | 58.3 | 56.9 | 59.8 | |
| ParaphraseTraining=Fine-tuning2023.10 | 57.8 | 57.5 | 58.2 | |
| RINANTE2021.01 | 42.1 | 49.4 | 36.7 | |
| ChatGPTPrompting=100-shot In Context Learning (ICL)2023.10 | 38.6 | 36.7 | 40.7 |