Aspect-level Sentiment Analysis on Laptop (test)
86.68AccuracyFlan-T5 + RVISAg
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Flan-T5 + RVISAgModel Parameters=11B, Rationale=GPT-3.52024.07 | 86.68 | 84.05 | |
| Flan-T5 + RVISAvModel Parameters=11B, Rationale=Vicuna2024.07 | 86.52 | 83.28 | |
| BERT Asp + SCAPTModel Parameters=110M2024.07 | 82.76 | 79.15 | |
| T5 Base + ABSA-ESAModel Parameters=220M2024.07 | 82.44 | 79.34 | |
| Flan-T5 + THORModel Parameters=11B2024.07 | 82.29 | 78.78 | |
| Flan-T5 + promptModel Parameters=250M2024.07 | 81.98 | 77.93 | |
| Flan-T5 + promptModel Parameters=11B2024.07 | 81.82 | 77.69 | |
| BERT Asp + CEPTModel Parameters=110M2024.07 | 81.66 | 78.38 | |
| Flan-T5 + THORModel Parameters=250M2024.07 | 81.66 | 77.51 | |
| Flan-T5 + RVISAgModel Parameters=250M, Rationale=GPT-3.52024.07 | 81.19 | 77.13 | |
| Flan-T5 + RVISAvModel Parameters=250M, Rationale=Vicuna2024.07 | 80.72 | 76.49 | |
| BERT + ADA+Model Parameters=110M2024.07 | 78.96 | 74.18 | |
| BERT + SPC+Model Parameters=110M2024.07 | 78.22 | 73.45 | |
| BERT + RGATModel Parameters=110M2024.07 | 78.21 | 74.07 | |
| MGANModel=Full Model2018.11 | 76.21 | 71.42 | |
| MGAN w/o C2FAblation=Without Coarse-to-Fine module2018.11 | 74.8 | 69.63 | |
| MGAN w/o PIAblation=Without Position Information2018.11 | 72.98 | 67.71 | |
| IAN2018.11 | 72.1 | — | |
| RAM2018.11 | 72.08 | 68.43 | |
| TNDescription=Target Network (proposed base model)2018.11 | 70.58 | 65.34 | |
| MemNet2018.11 | 70.33 | 64.09 | |
| AE-LSTM2018.11 | 68.97 | 62.5 | |
| ATAE-LSTM2018.11 | 68.65 | 62.45 | |
| TD-LSTM2018.11 | 68.18 | 62.28 |