All Relationships

Total Relationships: 103278

Source Entity Source Type Relationship Target Entity Type Paper ID
RoBERTa material isBasedOn Masked Language Model method 8283.0
RoBERTa material isBasedOn Neural Network method 8691.0
RoBERTa material learnsFrom InductiveBiases otherentity 31268.0
RoBERTa material mapsTo Embedding Vector material 28383.0
RoBERTa material measuredBy Classification Accuracy metric 29974.0
RoBERTa material measuredBy UUAS metric 21861.0
RoBERTa material outperforms Shallow Learned Models method 24734.0
RoBERTa material partOf CrossLingualEmbeddings method 3949.0
RoBERTa material reduces Calibration Error metric 31264.0
RoBERTa material testedOn Main Test Set otherentity 24692.0
RoBERTa material testedOn Nahj al Balagha Test Set material 24692.0
RoBERTa material testedWith Textual Perturbation task 32495.0
RoBERTa material trainedOn Dataset material 29659.0
RoBERTa material trainedOn English material 21115.0
RoBERTa material trainedOn Language Acquisition Data material 41749.0
RoBERTa material trainedOn Pretraining Data material 31980.0
RoBERTa material trainedOn Samples otherentity 26915.0
RoBERTa material usedFor CoreferenceResolution task 31328.0
RoBERTa material usedFor Grammatical Errors otherentity 37831.0
RoBERTa material usedFor Predicting Editorial Change task 21707.0