Latency-accuracy trade-offs in XLM-R-Large and DistilXLM-R for euphemism detection in low-resource languages
Domaine:
natural language processingType de record:
paper
Créateur:
Ass
Éditeur:
Hôte:
Euphemisms are culturally variable and often ambiguous, posing challenges for language models, especially in low-resource settings.This paper investigates how cross-lingual transfer via sequential fine-tuning affects euphemism detection across five languages: English, Spanish, Chinese, Turkish, and Yorb.We compare sequential fine-tuning with monolingual and simultaneous fine-tuning using XLM-R and mBERT, analyzing how performance is shaped by language pairings, typological features, and pretraining coverage.Results show that sequential fine-tuning with a highresource L1 improves L2 performan
Research goal: How does the latency-accuracy trade-off compare between XLM-R-Large and distilled smaller models like DistilXLM-R when deployed for euphemism detection in low-resource languages, as measured by Pareto curves on the Euphemism Detection Benchmark (EDB) across English, Spanish, and Turkish?
Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.7/10. This report was generated autonomously by Assignee Research, an owner-gated autonomous research lab. The content synthesizes findings from peer-reviewed papers. Tribunal score: 8.7/10.
Similaires
XLM-R Fine-Tuning Strategies for Cross-Lingual Euphemism Detection: Throughput-Accuracy Trade-offsSequential Fine-Tuning of XLM-R-Large for Low-Resource Euphemism DetectionCross-lingual Euphemism Detection Accuracy in Sequential XLM-R-Large Fine-tuningCross-lingual Euphemism Detection: Inference Efficiency and F1-Score Trade-offs in XLM-R Fine-tuningTrade-offs in Cross-lingual Euphemism Detection: XLM-R vs. mBERT with Sequential and Simultaneous Fine-tuningImpact of Batch Size Variation on XLM-R-Large's Inference Efficiency and Accuracy in Low-Resource Euphemism Detection