List your company
ResearchFormerly WizWand

Sentiment Classification on SST

88.19AccuracySCD-RoBERTa_base
77.197280.051182.90585.7589Mar 29, 2018Nov 25, 2018Jul 24, 2019Mar 21, 2020Nov 17, 2020Jul 16, 2021Mar 15, 2022
Updated 8mo ago

Evaluation results

MethodLinks
SCD-RoBERTa_baseBackbone=RoBERTa-base2022.0388.19
SimCSE-RoBERTa_baseBackbone=RoBERTa-base2022.0387.26
USE_T+CNN (lrn w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=CNN2018.0387.21
RoBERTa [CLS]-embeddingBackbone=RoBERTa-base, Pooling=CLS2022.0386.71
USE_T+CNN (w2v w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=CNN2018.0386.69
USE_T+DAN (w2v w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=DAN2018.0386.62
BERT [CLS]-embeddingBackbone=BERT-base, Pooling=CLS2022.0386.49
USE_T+DAN (lrn w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=DAN2018.0386.24
IS-BERT_baseBackbone=BERT-base2022.0385.96
SCD-BERT_baseBackbone=BERT-base2022.0385.89
USE TUniversal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=None, Transfer Task Model=None2018.0385.38
USE_D+CNN (w2v w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=CNN2018.0385.29
USE_D+CNN (lrn w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=CNN2018.0385.27
CNN (lrn w.e.)Transfer Level=None, Embedding Initialization=learned on transfer task, Transfer Task Model=CNN2018.0384.9
SimCSE-BERT_baseBackbone=BERT-base2022.0384.57
Avg. BERT embeddingsBackbone=BERT-base, Pooling=Avg2022.0384.4
CNN (w2v w.e.)Transfer Level=Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=CNN2018.0383.74
USE_D+DAN (lrn w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=DAN2018.0383.41
USE_D+DAN (w2v w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=DAN2018.0382.14
Skip-thoughtModel=Skip-thought2022.0382
DAN (lrn w.e.)Transfer Level=None, Embedding Initialization=learned on transfer task, Transfer Task Model=DAN2018.0381.52
DAN (w2v w.e.)Transfer Level=Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=DAN2018.0380.24
GloVe embeddings (avg.)Model=GloVe (avg.)2022.0380.18
USE DUniversal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=None, Transfer Task Model=None2018.0377.62