List your company
ResearchFormerly WizWand

Word Prediction on LAMBADA CONTROL (all)

36AccuracyHuman
-1.448.281827.72Oct 26, 2016
Updated 8mo ago

Evaluation results

MethodLinks
Humanestimation_source=100 randomly-sampled CONTROL instances2016.1036
LSTMbaseline_category=Paperno et al. (2016)2016.1021.9
n-grambaseline_category=Paperno et al. (2016)2016.1019.1
n-gram + cachebaseline_category=Paperno et al. (2016)2016.1019.1
GA Reader + featuresfeatures=Wang et al. (2016)2016.109.3
GA Reader2016.108.8
AS Reader + featuresfeatures=Wang et al. (2016)2016.108.6
Memory networkbaseline_category=Paperno et al. (2016)2016.108.5
AS Reader2016.108.5
Modified Stanford Reader2016.107.4
Stanford Reader2016.107
LSTMbaseline_category=Context-restricted language model2016.102.4
n-grambaseline_category=Context-restricted language model2016.102.2
n-gram + cachebaseline_category=Context-restricted language model2016.102.2
Lastbaseline_category=Context-restricted non-stopword2016.100.9
Most frequentbaseline_category=Context-restricted non-stopword2016.100.4
Randombaseline_category=Context-restricted non-stopword2016.100.3
Firstbaseline_category=Context-restricted non-stopword2016.100.1
Random in contextbaseline_category=Paperno et al. (2016)2016.100
Random cap. in contextbaseline_category=Paperno et al. (2016)2016.100