Try a synonym, or sign up to suggest a template from the app.
Neural Network Architectures Used in Generative Modeling
Description
This template compares three neural network architectures used to generate molecules: variational autoencoders, which compress a real molecule through an encoder into a latent space and reconstruct it with a decoder, and generative adversarial networks, where a generator creates molecules from Gaussian noise that a discriminator classifies as real or fake. It also shows normalizing flow, which maps a real molecule into latent space and back using flow and inverse flow transformations.
BioRender icons and templates meet journal, grant, and copyright standards. For AI-generated figures, policies vary. Learn More
Don’t start from scratch.
Create professional scientific illustrations quickly and easily, even without any design expertise
- Get started with a huge library of editable icons and templates such as common biological pathways, anatomy, or genetics.
- Create figures that effectively communicate your research in half the time using our editable icons.
- Use our PDB tool to quickly generate and customize protein structures.
.gif)
