Aspect-Based Sentiment Analysis on Restaurant Full (test)
91.52AccuracyFlan-T5 + RVISAg
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Flan-T5 + RVISAgModel Parameters=11B, Base Model=Flan-T5, Prompting Strategy=RVISA, Rationale Generation Model=GPT-3.5-TURBO2024.07 | 91.52 | 86.85 | |
| Flan-T5 + RVISAvModel Parameters=11B, Base Model=Flan-T5, Prompting Strategy=RVISA, Rationale Generation Model=Vicuna-13B2024.07 | 91.25 | 86.57 | |
| Flan-T5 + promptModel Parameters=11B, Base Model=Flan-T52024.07 | 89.29 | 83.68 | |
| BERT Asp + SCAPTModel Parameters=110M, Base Model=BERT2024.07 | 89.11 | 83.79 | |
| Flan-T5 + THORModel Parameters=11B, Base Model=Flan-T5, Prompting Strategy=THOR2024.07 | 88.57 | 82.93 | |
| T5 Base + ABSA-ESAModel Parameters=220M, Base Model=T5 Base2024.07 | 88.29 | 81.74 | |
| Flan-T5 + THORModel Parameters=250M, Base Model=Flan-T5, Prompting Strategy=THOR2024.07 | 87.68 | 81.1 | |
| BERT Asp + CEPTModel Parameters=110M, Base Model=BERT2024.07 | 87.5 | 82.07 | |
| BERT + ADA+Model Parameters=110M, Base Model=BERT2024.07 | 87.14 | 80.05 | |
| Flan-T5 + promptModel Parameters=250M, Base Model=Flan-T52024.07 | 86.88 | 79.78 | |
| Flan-T5 + RVISAgModel Parameters=250M, Base Model=Flan-T5, Prompting Strategy=RVISA, Rationale Generation Model=GPT-3.5-TURBO2024.07 | 86.61 | 78.92 | |
| BERT + RGATModel Parameters=110M, Base Model=BERT2024.07 | 86.6 | 81.35 | |
| Flan-T5 + RVISAvModel Parameters=250M, Base Model=Flan-T5, Prompting Strategy=RVISA, Rationale Generation Model=Vicuna-13B2024.07 | 86.43 | 78.49 | |
| BERT + SPC+Model Parameters=110M, Base Model=BERT2024.07 | 83.57 | 77.16 |