Contrastive Learning for Cross-modal Artist Retrieval

August 12, 2023 Β· Declared Dead Β· πŸ› International Society for Music Information Retrieval Conference

πŸ‘» CAUSE OF DEATH: Ghosted
No code link whatsoever

"No code URL or promise found in abstract"

Evidence collected by the PWNC Scanner

Authors Andres Ferraro, Jaehun Kim, Sergio Oramas, Andreas Ehmann, Fabien Gouyon arXiv ID 2308.06556 Category cs.IR: Information Retrieval Citations 5 Venue International Society for Music Information Retrieval Conference Last Checked 4 months ago
Abstract
Music retrieval and recommendation applications often rely on content features encoded as embeddings, which provide vector representations of items in a music dataset. Numerous complementary embeddings can be derived from processing items originally represented in several modalities, e.g., audio signals, user interaction data, or editorial data. However, data of any given modality might not be available for all items in any music dataset. In this work, we propose a method based on contrastive learning to combine embeddings from multiple modalities and explore the impact of the presence or absence of embeddings from diverse modalities in an artist similarity task. Experiments on two datasets suggest that our contrastive method outperforms single-modality embeddings and baseline algorithms for combining modalities, both in terms of artist retrieval accuracy and coverage. Improvements with respect to other methods are particularly significant for less popular query artists. We demonstrate our method successfully combines complementary information from diverse modalities, and is more robust to missing modality data (i.e., it better handles the retrieval of artists with different modality embeddings than the query artist's).
Community shame:
Not yet rated
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

πŸ“œ Similar Papers

In the same crypt β€” Information Retrieval

Died the same way β€” πŸ‘» Ghosted