Angular-Based Word Meta-Embedding Learning
This is an incremental improvement for natural language processing tasks, addressing the challenge of unknown embedding quality in unsupervised meta-embedding learning.
This work tackled the problem of improving word embeddings by comparing meta-embeddings trained with angular-based loss functions versus standard distance-based losses, showing that cosine and KL-divergence objectives outperform existing methods on word similarity and relatedness datasets.
Ensembling word embeddings to improve distributed word representations has shown good success for natural language processing tasks in recent years. These approaches either carry out straightforward mathematical operations over a set of vectors or use unsupervised learning to find a lower-dimensional representation. This work compares meta-embeddings trained for different losses, namely loss functions that account for angular distance between the reconstructed embedding and the target and those that account normalized distances based on the vector length. We argue that meta-embeddings are better to treat the ensemble set equally in unsupervised learning as the respective quality of each embedding is unknown for upstream tasks prior to meta-embedding. We show that normalization methods that account for this such as cosine and KL-divergence objectives outperform meta-embedding trained on standard $\ell_1$ and $\ell_2$ loss on \textit{defacto} word similarity and relatedness datasets and find it outperforms existing meta-learning strategies.