List your company
ResearchFormerly WizWand

Question Classification on TREC

98.07AccuracyUSE_T+CNN (w2v w.e.)
89.053291.394193.73596.0759Aug 25, 2014Jul 6, 2016May 19, 2018Mar 30, 2020Feb 10, 2022Dec 23, 2023Nov 4, 2025
Updated 4mo ago

Evaluation results

MethodLinks
USE_T+CNN (w2v w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=CNN2018.0398.07
USE_D+CNN (lrn w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=CNN2018.0397.71
USE_D+CNN (w2v w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=CNN2018.0397.67
BAE: BERT2021.0497.6
USE_T+CNN (lrn w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=CNN2018.0397.44
CNN (w2v w.e.)Transfer Level=Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=CNN2018.0397.32
CNN+MCFAAdditional Context=Translation, N=ensemble2018.0696.8
MUPPET2022.1296.8
ROBERTa-CLTraining supervision type=clean labels, Method categorization=Fully-supervised2020.1096.68
USE_T+DAN (lrn w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=DAN2018.0396.6
CNN+B1Additional Context=Translation, N=ensemble2018.0696.4
CNN+B2Additional Context=Translation, N=ensemble2018.0696.4
BLSTM-2DCNNNN category=ours2016.1196.1
CNN+MCFAAdditional Context=Translation, N=102018.0696
TBCNNNN category=CNN2016.1196
USE_D+DAN (lrn w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=learned on transfer task, Transfer Task Model=DAN2018.0395.86
CNN (lrn w.e.)Transfer Level=None, Embedding Initialization=learned on transfer task, Transfer Task Model=CNN2018.0395.82
DSCNN2018.0695.6
USE_T+DAN (w2v w.e.)Universal Encoder=Transformer (USE_T), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=DAN2018.0395.51
Dep-CNN2018.0695.4
CNN+B1Additional Context=Translation, N=12018.0695.4
CNN+B2Additional Context=Translation, N=12018.0695.4
CNN+MCFAAdditional Context=Translation, N=12018.0695.4
DSCNNNN category=Other2016.1195.4
CNN+B2Additional Context=Translation, N=102018.0695.2
FADS-ICLBackbone=Llama-1 30B, Shots=128-shots2024.0595.2
FADS-ICLBackbone=Llama-2 70B, Shots=128-shots2024.0595.2
SVMs2014.0895
CNN+B1Additional Context=Translation, N=102018.0695
FADS-ICLBackbone=Llama-2 13B, Shots=128-shots2024.0595
BLSTM-2DPoolingNN category=ours2016.1194.8
USE_D+DAN (w2v w.e.)Universal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence & Word, Embedding Initialization=word2vec skip-gram, Transfer Task Model=DAN2018.0394.72
C-LSTMNN category=Other2016.1194.6
TopCNNAdditional Context=Topic, Granularity=ensemble2018.0694
DAN (lrn w.e.)Transfer Level=None, Embedding Initialization=learned on transfer task, Transfer Task Model=DAN2018.0393.88
BLSTM-AttNN category=ours2016.1193.8
FADS-ICLBackbone=Llama-1 13B, Shots=128-shots2024.0593.8
CNN-non-staticword vectors=fine-tuned pre-trained (word2vec)2014.0893.6
CNN2018.0693.6
CNN2015.0493.6
CNN-non-staticNN category=CNN2016.1193.6
CNN-400Emb.Size=4002019.0593.6
FADS-ICLBackbone=Llama-1 7B, Shots=128-shots2024.0593.6
STM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=CNN2019.0593.48
DCNN2014.0893
DCNN2015.0493
DCNNNN category=CNN2016.1193
BLSTMNN category=ours2016.1193
FADS-ICLBackbone=Llama-2 7B, Shots=128-shots2024.0593
STM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=LSTM2019.0592.96
CNN-staticword vectors=static pre-trained (word2vec)2014.0892.8
NLM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=CNN2019.0592.8
Avg. BERT embeddingsBackbone=BERT-base, Pooling=Avg2022.0392.8
STM+TSEDEmb.Size=50, Teacher Model Architecture=CNN2019.0592.76
LM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=CNN2019.0592.58
LM+TSEDEmb.Size=50, Teacher Model Architecture=LSTM2019.0592.56
USE TUniversal Encoder=Transformer (USE_T), Transfer Level=Sentence, Embedding Initialization=None, Transfer Task Model=None2018.0392.51
TopCNNAdditional Context=Topic, Granularity=word2018.0692.5
LM+TSEDEmb.Size=50, Teacher Model Architecture=CNN2019.0592.48
AdaSent2018.0692.4
AdaSent2015.0492.4
AdaSentNN category=Other2016.1192.4
NLM+TSEDEmb.Size=50, Teacher Model Architecture=CNN2019.0592.36
STM+TSEDEmb.Size=50, Teacher Model Architecture=LSTM2019.0592.36
CNN-multichannelword vectors=pre-trained (word2vec), channels=22014.0892.2
SkipThoughtData Source Type=Ordered Sentences2016.0292.2
combine-skipNN category=Other2016.1192.2
Skip-thoughtModel=Skip-thought2022.0392.2
NLM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=LSTM2019.0592.08
TopCNNAdditional Context=Topic, Granularity=sentence2018.0692
CNN-MCNN category=CNN2016.1192
LSTM-400Emb.Size=4002019.0592
LM+TSED+PT+2LEmb.Size=50, Teacher Model Architecture=LSTM2019.0591.91
P.V.2015.0491.8
NLM+TSEDEmb.Size=50, Teacher Model Architecture=LSTM2019.0591.8
CNN-AnaNN category=CNN2016.1191.37
CNN-randword vectors=randomly initialized2014.0891.2
FADS-ICLNumber of training shots (m)=64, LLM scale=1.5B2024.0591.2
USE DUniversal Encoder=Deep Averaging Network (USE_D), Transfer Level=Sentence, Embedding Initialization=None, Transfer Task Model=None2018.0391.19
ColD-Fusion2022.1291.04
BRNN2015.0491
CNN-50Emb.Size=502019.0591
BERT [CLS]-embeddingBackbone=BERT-base, Pooling=CLS2022.0391
FADS-ICLcandidate pool size (m)=2562024.0591
ENCEmb.Size=502019.0590.6
kNN-promptingNumber of training shots (m)=128, LLM scale=1.5B2024.0590.5
FADS-ICLNumber of training shots (m)=128, LLM scale=1.5B2024.0590.5
FADS-ICLBackbone=GPT-2 1.5B, Shots=128-shots2024.0590.5
FADS-ICLcandidate pool size (m)=1282024.0590.5
ConMeZOBackbone=RoBERTa-Large, Evaluation Protocol=prompt-conditioned, Time Constraint=equal wall-clock time2025.1190.4
FADS-ICLBackbone=GPT-2 0.8B, Shots=128-shots2024.0590.3
RNN2015.0490.2
kNN-promptingBackbone=Llama-2 70B, Shots=128-shots2024.0590.2
kNN-promptBackbone=Llama-2 70B, Shots=128-shots2024.0590.1
TM with ϕextension_type=words within a given cosine angle θ2021.0490.04
TM with kextension_type=k nearest words in embedding space2021.0489.82
SCD-BERT_baseBackbone=BERT-base2022.0389.8
LOZO-MBackbone=RoBERTa-Large, Evaluation Protocol=prompt-conditioned, Time Constraint=equal wall-clock time2025.1189.8
FP+Transformerfeature_extension=FP2021.0489.51
LOZOBackbone=RoBERTa-Large, Evaluation Protocol=prompt-conditioned, Time Constraint=equal wall-clock time2025.1189.4