Word Embedding based on Low-Rank Doubly Stochastic Matrix Decomposition
This work addresses a specific issue in natural language processing for researchers and practitioners, but it appears incremental as it builds on existing embedding methods with a novel optimization focus.
The paper tackles the problem of word embedding similarity not being optimized in learning by proposing a neighbor embedding method that learns an embedding simplex with minimal discrepancy to input neighborhoods, resulting in better performance for various queries compared to an existing approach.
Word embedding, which encodes words into vectors, is an important starting point in natural language processing and commonly used in many text-based machine learning tasks. However, in most current word embedding approaches, the similarity in embedding space is not optimized in the learning. In this paper we propose a novel neighbor embedding method which directly learns an embedding simplex where the similarities between the mapped words are optimal in terms of minimal discrepancy to the input neighborhoods. Our method is built upon two-step random walks between words via topics and thus able to better reveal the topics among the words. Experiment results indicate that our method, compared with another existing word embedding approach, is more favorable for various queries.