Multi-label Emotion Classification on ROCStories Plutchik emotion labels (test)
67.29PrecisionLEAM w/ BERT Features
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LEAM w/ BERT FeaturesModel Group=Adding Label Semantics, Sub-category=Label Embeddings, Embeddings=BERT-base-uncased2020.06 | 67.29 | 54.48 | 60.22 | |
| BERTModel Group=Baselines, Backbone=BERT-base-uncased, Fine-tuned=true2020.06 | 65.63 | 56.91 | 60.96 | |
| BERT + Labels as InputModel Group=Adding Label Semantics, Sub-category=Label Embeddings, Input=BERT with label sentences2020.06 | 63.05 | 61.7 | 62.36 | |
| LEAM w/ GloVeModel Group=Adding Label Semantics, Sub-category=Label Embeddings, Embeddings=GloVe (100d)2020.06 | 59.81 | 54.46 | 57.03 | |
| Paul and FrankModel Group=Baselines, Year=2019, Config=Self-attention architecture2020.06 | 59.66 | 51.33 | 55.18 | |
| Semi-supervisionModel Group=Adding Label Semantics, Sub-category=Label Correlation, Mechanism=Semi-supervised inference2020.06 | 57.94 | 76.35 | 65.88 | |
| Learned CorrelationsModel Group=Adding Label Semantics, Sub-category=Label Correlation, Mechanism=Tracking label-label correlations2020.06 | 56.5 | 71.47 | 63.11 | |
| BiLSTMModel Group=Baselines, Reference=Rashkin et al. (2018)2020.06 | 25.31 | 33.44 | 28.81 | |
| RENModel Group=Baselines, Reference=Rashkin et al. (2018), Full Name=Recurrent Entity Network2020.06 | 25.3 | 37.3 | 30.15 | |
| CNNModel Group=Baselines, Reference=Rashkin et al. (2018)2020.06 | 24.47 | 38.87 | 30.04 | |
| NPNModel Group=Baselines, Reference=Rashkin et al. (2018), Full Name=Neural Process Network2020.06 | 24.33 | 40.1 | 30.29 |