Deep Residual Mixture Models
arXiv:2006.12063
Abstract
We propose Deep Residual Mixture Models (DRMMs), a novel deep generative model architecture. Compared to other deep models, DRMMs allow more flexible conditional sampling: The model can be trained once with all variables, and then used for sampling with arbitrary combinations of conditioning variables, Gaussian priors, and (in)equality constraints. This provides new opportunities for interactive and exploratory machine learning, where one should minimize the user waiting for retraining a model. We demonstrate DRMMs in constrained multi-limb inverse kinematics and controllable generation of animations.
arXiv commentCode and examples can be found at https://github.com/PerttuHamalainen/DRMM
References in corpus (13)
- Auto-Encoding Variational Bayes
- Categorical Reparameterization with Gumbel-Softmax
- NICE: Non-linear Independent Components Estimation
- Variational Inference with Normalizing Flows
- The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
- A Neural Representation of Sketch Drawings
- Sum-Product Networks: A New Deep Architecture
- Improved Precision and Recall Metric for Assessing Generative Models
- Robust Motion In-betweening
- Generative Image Inpainting with Contextual Attention
- Low Bit-Rate Speech Coding with VQ-VAE and a WaveNet Decoder
- The power of deeper networks for expressing natural functions
- Sum-Product-Transform Networks: Exploiting Symmetries using Invertible Transformations