Generative Modeling

Topology Adaptive Snakes Improve Mask Generation for Image Inpainting

T-snake deformable models, which can segment complex-shaped structures, improve generated masks passed to a GAN for image inpainting

Short-Run MCMC Residual Network toward Energy Based Model

Non-convergent non-persistent short-run MCMC residual network beats VAE and GAN for image synthesis, interpolation, and reconstruction

Generative Recurrent Language Models with MCMC Inference

Utilizes image parsing to text description framework that generates text descriptions based on understanding of image and video content

Masters Thesis in Generative Modeling in Vision and Language

Thesis for Computer Science Masters program currently in progress under my advisor Dr. Song-Chun Zhu at UCLA

Formulation of Variational Lower Bound and Application of VAE to MNIST Dataset

Formulation of the evidence lower bound (ELBO) for the variational autoencoder and an application to synthesizing binary images