Justin Salamon
@justin_salamon
Head of Sound Design AI Research at Adobe. Machine learning and signal processing for audio & video. Musician. He/him.
ID: 2829968159
http://www.justinsalamon.com 14-10-2014 17:43:47
905 Tweet
2,2K Takipçi
771 Takip Edilen
How to best convert text metadata to captions with an LLM? Our ISMIR Conference 2024 paper digs into this question for music-text representation learning and proposes a new recipe: Augment, Drop, and Swap. Led by our former intern Ilaria Manco w/ Oriol (Uri) Nieto and me at Adobe Research