The authors show that standard text style-transfer metrics are unstable and manipulable, recommend BLEU against human rewrites as an additional benchmark, and report three architectures that improve on that metric.
Generating lyrics with variational autoencoder and multi-modal artist embeddings
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We present a system for generating song lyrics lines conditioned on the style of a specified artist. The system uses a variational autoencoder with artist embeddings. We propose the pre-training of artist embeddings with the representations learned by a CNN classifier, which is trained to predict artists based on MEL spectrograms of their song clips. This work is the first step towards combining audio and text modalities of songs for generating lyrics conditioned on the artist's style. Our preliminary results suggest that there is a benefit in initializing artists' embeddings with the representations learned by a spectrogram classifier.
fields
cs.CL 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Style Transfer for Texts: Retrain, Report Errors, Compare with Rewrites
The authors show that standard text style-transfer metrics are unstable and manipulable, recommend BLEU against human rewrites as an additional benchmark, and report three architectures that improve on that metric.