Papers with multi-modal encoder-decoder model
Data-Efficient Playlist Captioning With Musical and Linguistic Knowledge (2022.emnlp-main)
Copied to clipboard
| Challenge: | Music streaming services feature billions of playlists created by users, professional editors or algorithms. |
| Approach: | They propose a multi-modal encoder-decoder model for automatic playlist captioning that leverages linguistic and musical knowledge to generate correct and thematic captions. |
| Outcome: | The proposed model yields 2x-3x higher BLEU@4 and CIDEr than state-of-the-art captioning algorithms on a new playlists dataset from two major streaming services. |