4 Hits in 3.8 sec

GLSR-VAE: Geodesic Latent Space Regularization for Variational AutoEncoder Architectures [article]

Gaëtan Hadjeres and Frank Nielsen and François Pachet
2017 arXiv   pre-print
We propose in this paper GLSR-VAE, a Geodesic Latent Space Regularization for the Variational AutoEncoder architecture and its generalizations which allows a fine control on the embedding of the data into  ...  VAEs (Variational AutoEncoders) have proved to be powerful in the context of density modeling and have been used in a variety of contexts for creative purposes.  ...  We define the Geodesic Latent Space Regularization for the Variational Auto-Encoder (GLSR-VAE) by R geo (z; {g k }, θ) := K k=1 R k (z; θ) (5) where R k (z; θ) = log r k ∂G k ∂z k (z) . (6) The distributions  ... 
arXiv:1707.04588v1 fatcat:mrghhxj3sfaajjxtg3z2n23qoq

Attribute-based Regularization of Latent Spaces for Variational Auto-Encoders [article]

Ashis Pati, Alexander Lerch
2020 arXiv   pre-print
In this paper, we present a novel method to structure the latent space of a Variational Auto-Encoder (VAE) to encode different continuous-valued attributes explicitly.  ...  Consequently, post-training, the model can be used to manipulate the attribute by simply changing the latent code of the corresponding regularized dimension.  ...  Acknowledgements The authors would like to thank Nvidia Corporation for their donation of a Titan V awarded as part of the GPU (Graphics Processing Unit) grant program which was used for running several  ... 
arXiv:2004.05485v3 fatcat:am3cuewchfes7oqve7gvb2ux6i

Deep Learning Techniques for Music Generation – A Survey [article]

Jean-Pierre Briot, Gaëtan Hadjeres, François-David Pachet
2019 arXiv   pre-print
Architecture - What type(s) of deep neural network is (are) to be used? Examples are: feedforward network, recurrent network, autoencoder or generative adversarial networks.  ...  . - For what destination and for what use? To be performed by a human(s) (in the case of a musical score), or by a machine (in the case of an audio file).  ...  generation, named geodesic latent space regularization (GLSR), with a system named GLSR-VAE.  ... 
arXiv:1709.01620v4 fatcat:hma4znleorfpvh62cpupxu4fq4

A Comprehensive Survey on Deep Music Generation: Multi-level Representations, Algorithms, Evaluations, and Future Directions [article]

Shulei Ji, Jing Luo, Xinyu Yang
2020 arXiv   pre-print
In addition, we summarize the datasets suitable for diverse tasks, discuss the music representations, the evaluation methods as well as the challenges under different levels, and finally point out several  ...  [143] proposed GLSR-VAE architecture to control data embedding in the latent space.  ...  First determine the geometric structure of the latent space, and then use the geodesic potential space regularization (GLSR) method to increase the loss of VAE.  ... 
arXiv:2011.06801v1 fatcat:cixou3d2jzertlcpb7kb5x5ery