Context-predicting models obtain a thorough and resounding victory against count-based counterparts

Evidence

Glove: Global Vectors for Word Representation (ResearchGate (citing Baroni, Dinu & Kruszewski))

Categorization

• corroborated: 2 independent passages support this claim
• XREF: connects to: Training word vectors solely to predict context words yields powerful meaning representations; Count-based and prediction-based methods are not fundamentally different, as both exploit the co-occurrence matrix — GloVe: Global Vectors for Word Representation; AI’s apparent generality exceeds what headline metrics justify — , The Braided Reading EPUB: 5-Year…, 1783

Related

• Count-based and prediction-based methods are not fundamentally different, as both exploit the co-occurrence matrix (similarity 0.94)
• Training word vectors solely to predict context words yields powerful meaning representations (similarity 0.79)