Papers with enhanced approach
Quantifying Context Overlap for Training Word Embeddings (D18-1)
Copied to clipboard
| Challenge: | Experimental results show that word embeddings can be improved using word embeds . word embedings are a popular form of natural language processing . |
| Approach: | They propose to estimate second order co-occurrence relations based on context overlap . they use the augmented data to enhance word embeddings learning . |
| Outcome: | The proposed model improves word vectors for word similarity and downstream NLP tasks. |