Musical Word Embedding: Bridging the Gap between Listening Contexts and Music
This work addresses the gap in vocabulary size and musical relevance for word embeddings in music information retrieval, but it appears incremental as it builds on existing embedding techniques without introducing a new paradigm.
The paper tackled the problem of representing words for music information retrieval by training word embeddings using both general text and music-specific data, and evaluated how well these embeddings associate listening contexts with musical compositions.
Word embedding pioneered by Mikolov et al. is a staple technique for word representations in natural language processing (NLP) research which has also found popularity in music information retrieval tasks. Depending on the type of text data for word embedding, however, vocabulary size and the degree of musical pertinence can significantly vary. In this work, we (1) train the distributed representation of words using combinations of both general text data and music-specific data and (2) evaluate the system in terms of how they associate listening contexts with musical compositions.