List your company
ResearchFormerly WizWand

Question Retrieval on CQADupStack (dev)

0.163Average PrecisionSBERT-supervised
0.026760.062130.09750.13287Apr 14, 2021
Updated 8mo ago

Evaluation results

MethodLinks
SBERT-supervisedMethod Category=In-domain supervised (Upper bound)2021.040.163
USE-largeMethod Category=Out-of-the-box supervised, Note=Trained with in-domain training data for this dataset2021.040.159
TSDAETraining Strategy=Unsupervised learning, Backbone=BERT-base2021.040.145
TSDAETraining Strategy=Domain adaptation, Protocol=target task → NLI+STS2021.040.144
MLMTraining Strategy=Domain adaptation, Protocol=target task → NLI+STS2021.040.143
CTTraining Strategy=Domain adaptation, Protocol=NLI+STS → target task2021.040.142
BERT-flowTraining Strategy=Domain adaptation, Protocol=NLI+STS → target task2021.040.139
SimCSETraining Strategy=Domain adaptation, Protocol=NLI+STS → target task2021.040.138
TSDAETraining Strategy=Domain adaptation, Protocol=NLI+STS → target task2021.040.136
CTTraining Strategy=Domain adaptation, Protocol=target task → NLI+STS2021.040.134
CTTraining Strategy=Unsupervised learning, Backbone=BERT-base2021.040.133
BM25Method Category=Classical Information Retrieval2021.040.133
SimCSETraining Strategy=Domain adaptation, Protocol=target task → NLI+STS2021.040.131
SBERT-base-nli-stsb-v2Method Category=Out-of-the-box supervised2021.040.129
SimCSETraining Strategy=Unsupervised learning, Backbone=BERT-base2021.040.124
SBERT-base-nli-v2Method Category=Out-of-the-box supervised2021.040.118
MLMTraining Strategy=Unsupervised learning, Backbone=BERT-base2021.040.117
Avg. GloVeMethod Category=Unsupervised Approaches2021.040.1
MLMTraining Strategy=Domain adaptation, Protocol=NLI+STS → target task2021.040.097
BERT-flowTraining Strategy=Unsupervised learning, Backbone=BERT-base2021.040.092
BERT-base-uncasedMethod Category=Pre-trained2021.040.065
Sent2VecMethod Category=Unsupervised Approaches2021.040.032