Matches in SemOpenAlex for { <https://semopenalex.org/work/W4316658018> ?p ?o ?g. }
Showing items 1 to 66 of
66
with 100 items per page.
- W4316658018 abstract "Training semantic similarity model to detect duplicate text pairs is a challenging task as almost all of datasets are imbalanced, by data nature positive samples are fewer than negative samples, this issue can easily lead to model bias. Using traditional pairwise loss functions like pairwise binary cross entropy or Contrastive loss on imbalanced data may lead to model bias, however triplet loss showed improved performance compared to other loss functions. In triplet loss-based models data is fed to the model as follow: anchor sentence, positive sentence and negative sentence. The original data is permutated to follow the input structure. The default structure of training samples data is 363,861 training samples (90% of the data) distributed as 134,336 positive samples and 229,524 negative samples. The triplet structured data helped to generate much larger amount of balanced training samples 456,219. The test results showed higher accuracy and f1 scores in testing. We fine-tunned RoBERTa pre trained model using Triplet loss approach, testing showed better results. The best model scored 89.51 F1 score, and 91.45 Accuracy compared to 86.74 F1 score and 87.45 Accuracy in the second-best Contrastive loss-based BERT model." @default.
- W4316658018 created "2023-01-17" @default.
- W4316658018 creator A5026137726 @default.
- W4316658018 creator A5042863110 @default.
- W4316658018 date "2022-10-12" @default.
- W4316658018 modified "2023-09-27" @default.
- W4316658018 title "Improving The Performance of Semantic Text Similarity Tasks on Short Text Pairs" @default.
- W4316658018 cites W2933138175 @default.
- W4316658018 cites W3107640582 @default.
- W4316658018 cites W3176804172 @default.
- W4316658018 cites W3202538305 @default.
- W4316658018 doi "https://doi.org/10.1109/esolec54569.2022.10009072" @default.
- W4316658018 hasPublicationYear "2022" @default.
- W4316658018 type Work @default.
- W4316658018 citedByCount "0" @default.
- W4316658018 crossrefType "proceedings-article" @default.
- W4316658018 hasAuthorship W4316658018A5026137726 @default.
- W4316658018 hasAuthorship W4316658018A5042863110 @default.
- W4316658018 hasConcept C103278499 @default.
- W4316658018 hasConcept C115961682 @default.
- W4316658018 hasConcept C12267149 @default.
- W4316658018 hasConcept C130318100 @default.
- W4316658018 hasConcept C153180895 @default.
- W4316658018 hasConcept C154945302 @default.
- W4316658018 hasConcept C167981619 @default.
- W4316658018 hasConcept C16910744 @default.
- W4316658018 hasConcept C184898388 @default.
- W4316658018 hasConcept C199360897 @default.
- W4316658018 hasConcept C204321447 @default.
- W4316658018 hasConcept C2777530160 @default.
- W4316658018 hasConcept C2988416141 @default.
- W4316658018 hasConcept C41008148 @default.
- W4316658018 hasConcept C51632099 @default.
- W4316658018 hasConcept C66905080 @default.
- W4316658018 hasConceptScore W4316658018C103278499 @default.
- W4316658018 hasConceptScore W4316658018C115961682 @default.
- W4316658018 hasConceptScore W4316658018C12267149 @default.
- W4316658018 hasConceptScore W4316658018C130318100 @default.
- W4316658018 hasConceptScore W4316658018C153180895 @default.
- W4316658018 hasConceptScore W4316658018C154945302 @default.
- W4316658018 hasConceptScore W4316658018C167981619 @default.
- W4316658018 hasConceptScore W4316658018C16910744 @default.
- W4316658018 hasConceptScore W4316658018C184898388 @default.
- W4316658018 hasConceptScore W4316658018C199360897 @default.
- W4316658018 hasConceptScore W4316658018C204321447 @default.
- W4316658018 hasConceptScore W4316658018C2777530160 @default.
- W4316658018 hasConceptScore W4316658018C2988416141 @default.
- W4316658018 hasConceptScore W4316658018C41008148 @default.
- W4316658018 hasConceptScore W4316658018C51632099 @default.
- W4316658018 hasConceptScore W4316658018C66905080 @default.
- W4316658018 hasLocation W43166580181 @default.
- W4316658018 hasOpenAccess W4316658018 @default.
- W4316658018 hasPrimaryLocation W43166580181 @default.
- W4316658018 hasRelatedWork W1997312918 @default.
- W4316658018 hasRelatedWork W2038246283 @default.
- W4316658018 hasRelatedWork W2047828095 @default.
- W4316658018 hasRelatedWork W2116838603 @default.
- W4316658018 hasRelatedWork W2252122760 @default.
- W4316658018 hasRelatedWork W2365659184 @default.
- W4316658018 hasRelatedWork W2766760871 @default.
- W4316658018 hasRelatedWork W2899468685 @default.
- W4316658018 hasRelatedWork W3078371441 @default.
- W4316658018 hasRelatedWork W78638240 @default.
- W4316658018 isParatext "false" @default.
- W4316658018 isRetracted "false" @default.
- W4316658018 workType "article" @default.