List your company
ResearchFormerly WizWand

Top-level Implicit Discourse Relation Recognition on PDTB Lin split 2.0

72.78F1DiscoPrompt (T5-11b)
23.17236.05148.9361.809May 6, 2023
Updated 8mo ago

Evaluation results

MethodLinks
DiscoPrompt (T5-11b)Backbone=T5-11b2023.0572.7877.55
DiscoPromptBackbone=T5, Model Scale=11b2023.0572.7877.55
DiscoPrompt (T5-large)Backbone=T5-large2023.0567.0673.76
DiscoPromptBackbone=T5, Model Scale=large2023.0567.0673.76
DiscoPrompt (T5-base)Backbone=T5-base2023.0564.971.28
DiscoPromptBackbone=T5, Model Scale=base2023.0564.971.28
Prompt-Tuning (T5-large)Backbone=T5-large, Training Strategy=Prompt-Tuning2023.0559.9271.02
Prompt-TuningBackbone=T5, Model Scale=large2023.0559.9271.02
XLNet (large, cased)Backbone=XLNet-large2023.0558.9772.17
XLNetBackbone=XLNet, Model Scale=large, Casing=cased2023.0558.9772.17
Liu et al.2023.0558.5468.66
BMGF-ROBERTaBackbone=RoBERTa2023.0558.5468.66
Fine-Tuning (T5-large)Backbone=T5-large, Training Strategy=Fine-Tuning2023.0558.1271.13
Fine-TuningBackbone=T5, Model Scale=large2023.0558.1271.13
Prefix-Tuning (T5-large)Backbone=T5-large, Training Strategy=Prefix-Tuning2023.0558.0669.84
Prefix-TuningBackbone=T5, Model Scale=large2023.0558.0669.84
XLNet(base, cased)Backbone=XLNet-base2023.0556.1668.05
XLNetBackbone=XLNet, Model Scale=base, Casing=cased2023.0556.1668.05
BERT-largeBackbone=BERT-large2023.0556.0668.4
PCP w/ ROBERTa-baseBackbone=RoBERTa-base2023.055366.58
PCPBackbone=RoBERTa-base2023.055366.58
PCP w/ ROBERTa-largeBackbone=RoBERTa-large2023.0552.7571.13
PCPBackbone=RoBERTa-large2023.0552.7571.13
He et al.2023.0551.2459.94
TransS2023.0551.2459.94
Fine-Tuning (T5-base)Backbone=T5-base, Training Strategy=Fine-Tuning2023.0550.563.59
Fine-TuningBackbone=T5, Model Scale=base2023.0550.563.59
BERT-baseBackbone=BERT-base2023.0543.4463.46
Prompt-Tuning (T5-base)Backbone=T5-base, Training Strategy=Prompt-Tuning2023.0525.2655.09
Prompt-TuningBackbone=T5, Model Scale=base2023.0525.2655.09
Prefix-Tuning (T5-base)Backbone=T5-base, Training Strategy=Prefix-Tuning2023.0525.0854.18
Prefix-TuningBackbone=T5, Model Scale=base2023.0525.0854.18