Deep Generative Models and Biological Applications

<p>High-dimensional probability distributions are important objects in a wide variety of applications. </p><p>Generative models provide an excellent manipulation method for training from rich available unlabeled data set and sampling new data points from underlying high-dimensional probability distributions. </p><p>The recent proposed Variational auto-encoders (VAE) framework is an efficient high-dimensional inference method to modeling complicated data manifold in an approximate Bayesian way, i.e., variational inference. </p><p>We first discuss how to design fast stochastic backpropagation algorithm for the VAE based amortized variational inference method. </p><p>Particularly, we propose second order Hessian-free optimization method for Gaussian latent variable models and provide a theoretical justification to the convergence of Monte Carlo estimation in our algorithm. </p><p>Then, we apply the amortized variational inference to a dynamic modeling application in flu diffusion task. </p><p>Compared with traditional approximate Gibbs sampling algorithm, we make less assumption to the infection rate. </p><p>Differing from the maximum likelihood approach of VAE, Generative Adversarial Networks (GAN) is trying to solve the generation problem from a game theoretical way. </p><p>From this viewpoint, we design a framework VAE+GAN, by placing a discriminator on top of auto-encoders based model and introducing an extra adversarial loss. </p><p>The adversarial training induced by the classification loss is to make the discriminator believe the sample from the generative model is as real as the one from the true dataset. </p><p>This trick can practically improve the quality of generation samples, demonstrated on images and text domains with elaborately designed architectures. </p><p>Additionally, we validate the importance of generative adversarial loss with the conditional generative model in two biological applications: approximate Turing pattern PDEs generation in synthetic/system biology, and automatic cardiovascular disease detection in medical imaging processing.</p>

Author dc:creator

• Fan, Kai

Advisor dc:contributor.advisor

• Heller, Katherine

• Statistics
• Artificial intelligence
• Bioinformatics
• Adversarial training
• Deep Generative Models
• Fast inference
• Generative adversarial nets
• Variational auto-encoders

Handle dc:identifier.uri

https://hdl.handle.net/10161/16785

OAI identifier oai:identifier

oai:dukespace.lib.duke.edu:10161/16785