List your company
ResearchFormerly WizWand

Machine Comprehension on CBT-NE (val)

82.3AccuracyAS Reader (+BookTest, ensemble)
49.95658.35366.7575.147Mar 4, 2016Mar 19, 2016Apr 4, 2016Apr 20, 2016May 6, 2016May 22, 2016Jun 7, 2016
Updated 8mo ago

Evaluation results

MethodLinks
AS Reader (+BookTest, ensemble)trained_on_large_set=true, ensemble=true2016.0682.3
AS Reader (+BookTest)trained_on_large_set=true2016.0680.5
GA Readervariant=GA (+feature, fix L(w))2016.0678.5
NSE2016.0678.2
AoA Reader2016.0677.8
GA Readervariant=GA (fix L(w))2016.0677.2
GA Readervariant=GA (+feature, update L(w))2016.0677.2
Iterative Alternating Neural Attentionensemble=true2016.0676.9
Iterative Attentive Readerensemble=true2016.0676.9
GA Readervariant=GA (update L(w))2016.0676.7
EpiReaderensemble=true2016.0676.6
EpiReaderensemble=true2016.0676.6
AS ReaderModel variant=greedy ensemble2016.0376.2
AS Readerensemble=true2016.0676.2
EpiReader2016.0675.3
EpiReader2016.0675.3
Iterative Alternating Neural Attention2016.0675.2
Iterative Attentive Reader2016.0675.2
GA Readervariant=GA--2016.0674.9
AS Readerensemble=true2016.0674.5
AS ReaderModel variant=avg ensemble2016.0374.5
AS Readerensemble=true2016.0674.5
AS Reader2016.0673.8
AS ReaderModel variant=single model2016.0373.8
AS Reader2016.0673.8
AS Reader2016.0673.8
AS ReaderModel variant=avg for top 20%2016.0373.3
Iterative Alternating Neural Attentionattention=fixed query attention2016.0673.3
EpiReader Extractor2016.0673.2
MemNNs2016.0670.4
MemNNsModel variant=window memory + self-sup.2016.0370.4
MemNNsconfig=window memory + self-sup.2016.0670.4
MemNets2016.0670.4
LSTMscontext + query=true2016.0651.2
LSTMsInput configuration=context+query2016.0351.2
LSTMsinput=context+query2016.0651.2
LSTMS (context + query)2016.0651.2