Sentiment Classification on Yelp Polarity (test)
1.81Error RateBERT-ITPT-FiT
Evaluation Results
| Method | Links | |
|---|---|---|
| BERT-ITPT-FiTBackbone=BERT-Large, Further pre-training=In-Task (ITPT)2019.05 | 1.81 | |
| BERT-ITPT-FiTBackbone=BERT-Base, Further pre-training=In-Task (ITPT)2019.05 | 1.92 | |
| BERT-FiTBackbone=BERT-Large2019.05 | 2.04 | |
| ULMFiT2019.05 | 2.16 | |
| BERT-FiTBackbone=BERT-Base2019.05 | 2.28 | |
| M-ACNNNumber of convolutional filters=Multiple2017.09 | 3.89 | |
| Self-attentive Embedding2017.09 | 3.92 | |
| VDCNNDepth=29, Pooling=MaxPooling2016.06 | 4.28 | |
| Deep CNNNetwork depth=29 layers2017.09 | 4.28 | |
| VDCNNDepth=29, Pooling=Convolution2016.06 | 4.35 | |
| ngrams2015.09 | 4.36 | |
| ngramsAuthor=Zhang2016.06 | 4.36 | |
| ngramsType=bag-of-means, N-grams=up to 5-grams2017.09 | 4.36 | |
| VDCNNDepth=17, Pooling=MaxPooling2016.06 | 4.5 | |
| Deep CNNNetwork depth=17 layers2017.09 | 4.5 | |
| ngrams TFIDFTF-IDF Weighting=true2015.09 | 4.56 | |
| ngrams TFIDFType=TFIDF, N-grams=up to 5-grams2017.09 | 4.56 | |
| M-CNNNumber of convolutional filters=Multiple2017.09 | 4.58 | |
| VDCNNDepth=29, Pooling=KMaxPooling2016.06 | 4.63 | |
| Full Conv. (large, Th)Model Size=large, Case Sensitive=true, Thesaurus Augmentation=true2015.09 | 4.88 | |
| VDCNNDepth=9, Pooling=MaxPooling2016.06 | 4.88 | |
| Deep CNNNetwork depth=9 layers2017.09 | 4.88 | |
| Large word CNNScale=Large, Features per layer=1024, Total layers=62017.09 | 4.89 | |
| VDCNNDepth=17, Pooling=Convolution2016.06 | 4.96 | |
| VDCNNDepth=9, Pooling=Convolution2016.06 | 5.01 | |
| Lk. Conv. (large, Th)Model Size=large, Thesaurus Augmentation=true2015.09 | 5.03 | |
| VDCNNDepth=17, Pooling=KMaxPooling2016.06 | 5.05 | |
| LSTM2015.09 | 5.26 | |
| VDCNNDepth=9, Pooling=KMaxPooling2016.06 | 5.27 | |
| Full Conv. (small, Th)Model Size=small, Case Sensitive=true, Thesaurus Augmentation=true2015.09 | 5.42 | |
| Small word CNNScale=Small, Features per layer=256, Total layers=62017.09 | 5.54 | |
| Conv. (large, Th)Model Size=large, Thesaurus Augmentation=true2015.09 | 5.82 | |
| BoW TFIDFTF-IDF Weighting=true2015.09 | 6.34 | |
| S-ACNNNumber of convolutional filters=Single2017.09 | 6.41 | |
| BoW2015.09 | 7.76 | |
| Bag-of-means2015.09 | 12.67 | |
| S-CNNNumber of convolutional filters=Single2017.09 | 14.48 |