RoBERTa + SubRegWeigh (K-means)

Reported on 5 benchmarks across 2 tasks · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing5 results

Sentiment AnalysisonSST-2 Binary classification
Accuracy· 2024-09-10
94.84
best: 97.5 (T5-11B)
SubRegWeigh: Effective and Efficient Annotation Weighing with Subword Regularization arXiv:2409.06216
Named Entity Recognition (NER)onWNUT 2017
F1· 2024-09-10
60.29
best: 60.45 (CL-KL)
SubRegWeigh: Effective and Efficient Annotation Weighing with Subword Regularization arXiv:2409.06216
Named Entity Recognition (NER)onCoNLL-2020
F1· 2024-09-10
94.96
best: 95.31 (LUKE + SubRegWeigh (K-means))
SubRegWeigh: Effective and Efficient Annotation Weighing with Subword Regularization arXiv:2409.06216
Named Entity Recognition (NER)onCoNLL 2003 (English)
F1· 2024-09-10
93.81
best: 94.6 (ACE + document-context)
SubRegWeigh: Effective and Efficient Annotation Weighing with Subword Regularization arXiv:2409.06216
Named Entity Recognition (NER)onCoNLL++
F1· 2024-09-10
95.45
best: 96.12 (LUKE + SubRegWeigh (K-means))
SubRegWeigh: Effective and Efficient Annotation Weighing with Subword Regularization arXiv:2409.06216