List your company
ResearchFormerly WizWand

Machine Comprehension on CNN (val)

0.779AccuracyGA Reader
0.494040.568020.6420.71598Mar 4, 2016Apr 14, 2016May 25, 2016Jul 5, 2016Aug 15, 2016Sep 25, 2016Nov 5, 2016
Updated 8mo ago

Evaluation results

MethodLinks
GA Readervariant=GA (update L(w))2016.060.779
GA Readervariant=GA (fix L(w))2016.060.779
GA Readervariant=GA (+feature, update L(w))2016.060.773
Stanford ARensemble=true2016.110.772
Stanford AR (relabeling)ensemble=true2016.060.772
GA Readervariant=GA (+feature, fix L(w))2016.060.767
GAReaderensemble=true2016.110.764
BiDAFensemble=false2016.110.763
BIDAF2016.060.763
Iterative Alternating Neural AttentionEnsemble=true2016.060.752
Iterative Attentive Readerensemble=true2016.060.752
Iterative Attentionensemble=true2016.110.745
AS ReaderAggregation=greedy ensemble2016.030.745
AS Readerensemble=true2016.110.739
AS ReaderAggregation=avg ensemble2016.030.739
AS ReaderEnsemble=true2016.060.739
AS Readerensemble=true2016.060.739
Stanford ARensemble=false2016.110.738
Stanford AR (relabeling)2016.060.738
EpiReader2016.060.734
EpiReaderensemble=false2016.110.734
EpiReader2016.060.734
EpiReader2016.090.734
AoA Readerensemble=false2016.110.731
AoA Reader2016.060.731
AoA Reader2016.090.731
GAReaderensemble=false2016.110.73
GA Readervariant=GA--2016.060.73
GA Reader2016.090.73
ReasoNetensemble=false2016.110.729
ReasoNet2016.060.729
ReasoNetTmax=52016.090.729
Iterative Attentionensemble=false2016.110.726
Iterative Alternating Neural AttentionEnsemble=false2016.060.726
Iterative Attentive Reader2016.060.726
Iterative Attention Reader2016.090.726
Stanford AR2016.060.724
Chen et al. (2016)Model configuration=single model2016.030.724
Stanford AREmbeddings=GloVe2016.060.724
Stanford AR2016.090.722
EpiReader Extractor2016.060.718
DER Networkensemble=false2016.110.713
Dynamic Entity Repres.Strategy=+ w2v2016.030.713
DER Network2016.060.713
DER Network2016.090.713
Dynamic Entity Repres.Strategy=max-pool2016.030.712
Dynamic Entity Repres.Strategy=max-pool + byway2016.030.708
AS Reader2016.060.686
AS Readerensemble=false2016.110.686
AS ReaderModel configuration=single model2016.030.686
AS Reader2016.060.686
AS Reader2016.060.686
AS Reader2016.090.686
AS ReaderAggregation=avg for top 20%2016.030.684
MemNNensemble=true2016.110.662
MemNNsModel configuration=ensemble2016.030.662
MemNNsEnsemble=true2016.060.662
MemNetsensemble=true2016.060.662
MemNNs2016.060.634
MemNNensemble=false2016.110.634
MemNNsModel configuration=single model2016.030.634
MemNNs (window memory + self sup.)2016.060.634
MemNets2016.060.634
MemNets2016.090.634
Impatient Reader2016.060.618
Impatient ReaderModel type=single model2016.030.618
Impatient Reader2016.060.618
Impatient Reader2016.060.618
Attentive Reader2016.060.616
Attentive Readerensemble=false2016.110.616
Attentive ReaderModel type=single model2016.030.616
Attentive Reader2016.060.616
Attentive Reader2016.060.616
Attentive Reader2016.090.616
Deep LSTM Reader2016.060.55
Deep LSTM Reader2016.060.55
Deep LSTM Reader2016.060.55
Deep LSTM Reader2016.090.55
MemNNs (window memory)2016.060.508
Word distance model2016.060.505