New method improves Random Indexing embeddings with sparse PPMI graph averaging

Researchers have developed a method to improve Random Indexing (RI) embeddings by averaging them with a sparse Positive Pointwise Mutual Information (PPMI) graph. This technique, tested on a fairytales corpus, enhanced accuracy on a semantic analogy task from 19.4% to 30.7%. While this graph averaging method showed promise for RI, it performed less competitively compared to neural baselines like Skip-gram and CBOW on other datasets and tasks. AI

IMPACT Introduces a novel technique for refining word embeddings, potentially offering an alternative to gradient-based methods for specific NLP tasks.

RANK_REASON Academic paper detailing a novel method for improving word embeddings. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• continuous bag-of-words model
• Positive Pointwise Mutual Information
• PPMI+SVD
• Random Indexing
• SimLex-999
• singular value decomposition
• Skip-gram
• William Andreopoulos

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →