no code implementations • EACL (VarDial) 2021 • Kushagra Bhatia, Divyanshu Aggarwal, Ashwini Vaidya
In this paper we compare the performance of three models: SGNS (skip-gram negative sampling) and augmented versions of SVD (singular value decomposition) and PPMI (Positive Pointwise Mutual Information) on a word similarity task.