List your company
ResearchFormerly WizWand

Classification on bbc

99.72AccuracyMPAD-clique
64.651273.755682.8691.9644Aug 17, 2019Oct 2, 2020Nov 18, 2021Jan 4, 2023Feb 20, 2024Apr 7, 2025May 24, 2026
Updated 5mo ago

Evaluation results

MethodLinks
MPAD-clique2019.0899.72——
MPAD-path2019.0899.59——
MPAD-sentence-att2019.0899.32——
WMD2019.0898.71——
doc2vec2019.0898.64——
CNN2019.0898.37——
MPAD2019.0898.37——
MAGDIFFPre-processing=None, Classifier=5-NN2023.1198——
HN-ATT2019.0896.73——
DP-SGD-RCProj Dim. (k)=32, Epsilon (ϵ)=9, Fine-tuning protocol=full fine-tuning, Model=Llama-3.2-1B2026.0596.4——
DP-SGD (FGC)Proj Dim. (k)=N/A, Epsilon (ϵ)=9, Fine-tuning protocol=full fine-tuning, Model=Llama-3.2-1B2026.0596.33——
DiSAN2019.0896.05——
DP-SGD-RCProj Dim. (k)=32, Epsilon (ϵ)=2, Fine-tuning protocol=full fine-tuning, Model=Llama-3.2-1B2026.0595.6——
LSTM-GRNN2019.0895.52——
InfOptTraining Data Percentage=30% Data2026.0595.3——
non-privateProj Dim. (k)=N/A, Epsilon (ϵ)=N/A, Fine-tuning protocol=full fine-tuning, Model=Llama-3.2-1B2026.0595.2——
MAGDIFFPre-processing=PCA, Classifier=5-NN2023.1195——
CDVM5kTraining Data Percentage=30% Data2026.0595——
InfOptTraining Data Percentage=25% Data2026.0595——
SPGK2019.0894.97——
CDVM5kTraining Data Percentage=20% Data2026.0594.9——
BanzhafTraining Data Percentage=30% Data2026.0594.8——
CDVM10kTraining Data Percentage=25% Data2026.0594.7——
CDVM10kTraining Data Percentage=20% Data2026.0594.7——
CDVM10kTraining Data Percentage=15% Data2026.0594.7——
CDVM5kTraining Data Percentage=25% Data2026.0594.6——
CDVM5kTraining Data Percentage=15% Data2026.0594.6——
CDVM10kTraining Data Percentage=30% Data2026.0594.6——
DataOOBTraining Data Percentage=25% Data2026.0594.5——
BanzhafTraining Data Percentage=25% Data2026.0594.5——
InfOptTraining Data Percentage=20% Data2026.0594.4——
DataOOBTraining Data Percentage=30% Data2026.0594.4——
BanzhafTraining Data Percentage=20% Data2026.0594.4——
DAN2019.0894.3——
CDVM10kTraining Data Percentage=10% Data2026.0594.3——
RandomTraining Data Percentage=30% Data2026.0594.3——
CDVM5kTraining Data Percentage=10% Data2026.0594.1——
DP-SGD (FGC)Proj Dim. (k)=N/A, Epsilon (ϵ)=2, Fine-tuning protocol=full fine-tuning, Model=Llama-3.2-1B2026.0594.06——
DataOOBTraining Data Percentage=20% Data2026.0594——
RandomTraining Data Percentage=25% Data2026.0594——
InfOptTraining Data Percentage=10% Data2026.0593.8——
DataOOBTraining Data Percentage=15% Data2026.0593.8——
BanzhafTraining Data Percentage=15% Data2026.0593.6——
CDVM10kTraining Data Percentage=5% Data2026.0593.4——
InfOptTraining Data Percentage=15% Data2026.0593.4——
RandomTraining Data Percentage=20% Data2026.0593.4——
CDVM5kTraining Data Percentage=5% Data2026.0593.3——
RandomTraining Data Percentage=15% Data2026.0592.6——
DataOOBTraining Data Percentage=10% Data2026.0592.1——
DataOOBTraining Data Percentage=5% Data2026.0591.2——
BanzhafTraining Data Percentage=10% Data2026.0590.5——
RandomTraining Data Percentage=10% Data2026.0590.1——
RandomTraining Data Percentage=5% Data2026.0584.9——
VSPre-processing=None, Classifier=5-NN2023.1184——
VSPre-processing=PCA, Classifier=5-NN2023.1184——
InfOptTraining Data Percentage=5% Data2026.0583.6——
BanzhafTraining Data Percentage=5% Data2026.0580.1——
AVGSIMPre-processing=None, Classifier=5-NN2023.1174——
AVGSIMPre-processing=PCA, Classifier=5-NN2023.1173——
GMSTDSPre-processing=None, Classifier=5-NN2023.1166——
GMSTDSPre-processing=PCA, Classifier=5-NN2023.1166——
DP-SGD (FGC)Base Model=Llama3 1B, Fine-tuning protocol=LoRA, Proj Dim. (k)=N/A2026.05—96.393.1
DP-SGD-RCBase Model=Llama3 1B, Fine-tuning protocol=LoRA, Proj Dim. (k)=322026.05—95.993.2
non-privateBase Model=Llama3 1B, Fine-tuning protocol=LoRA, Proj Dim. (k)=N/A2026.05—96.5—