Aspect-based Sentiment Analysis on SemEval Restaurant 2014 (All)
87.45F1 ScoreTHOR
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| THORModel Parameters=11B, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning2023.05 | 87.45 | — | — | |
| THORModel Parameters=11B, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning, Method Variation=w/o Reason-revising2023.05 | 86.88 | — | — | |
| THORModel Parameters=11B, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning, Method Variation=w/o Self-consistency2023.05 | 86.03 | — | — | |
| Flan-T5+PromptModel Parameters=11B, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning2023.05 | 84.72 | — | — | |
| BERTAsp+SCAPTModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 83.79 | — | — | |
| THORModel Parameters=250M, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning2023.05 | 82.98 | — | — | |
| BERTAsp+CEPTModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 82.07 | — | — | |
| Flan-T5+PromptModel Parameters=250M, Backbone=Flan-T5, Evaluation Protocol=Supervised fine-tuning2023.05 | 81.5 | — | — | |
| BERT+ISAIVModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 81.4 | — | — | |
| BERT+RGATModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 81.35 | — | — | |
| BERT+PromptModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 81.34 | — | — | |
| BERT+ADAModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 80.05 | — | — | |
| BERT+SPCModel Parameters=110M, Backbone=BERT, Evaluation Protocol=Supervised fine-tuning2023.05 | 77.16 | — | — | |
| Extra2023.07 | 60.09 | 54.7 | 66.67 | |
| CoTAMshot_number (K)=10, generation_count (N)=3, backbone=RoBERTa-Large, embedding_model=SimCSE2023.07 | 56.36 | 50 | 64.58 | |
| FlipDA++2023.07 | 51.38 | 45.9 | 58.33 | |
| LLM SL2023.07 | 49.54 | 44.26 | 56.25 | |
| CoTDA2023.07 | 47.57 | 44.55 | 51.04 | |
| Base2023.07 | 34.82 | 30.61 | 40.38 |