Skip to content

Music Informatics Group

Georgia Tech Center for Music Technology

  • blog
  • publications
  • contact

Projects & Events

  • Score-Informed Networks for Music Performance Assessment
  • Multi-Task Learning for Instrument Activation Aware Music Source Separation
  • AR-VAE: Attribute-based Regularization of VAE Latent Spaces
  • dMelodies: A Music Dataset for Disentanglement Learning
  • Learning to Traverse Latent Spaces for Musical Score Inpainting
  • An Attention Mechanism for Musical Instrument Recognition
  • Explicitly Conditioned Melody Generation
  • Music Informatics Group @ICML Machine Learning for Music Discovery Workshop
  • From labeled to unlabeled data – on the data challenge in automatic drum transcription
  • On the evaluation of generative models in music
Previous Image
Next Image

disent_results_MIG

Figure 3: Disentanglement performance (higher is better) in terms of Mutual Information Gap (MIG) of different methods on the dMelodies and dSprites datasets.

Posted on August 12, 2020August 12, 2020Full size 3200 × 2400

Post navigation

Published indMelodies: A Music Dataset for Disentanglement Learning
Proudly powered by WordPress