Invited lecture at International AI Doctoral Academy (AIDA) AI Excellence Lecture Series

Please see the event information at: https://www.i-aida.org/events/generative-models-controllable-generation-and-learning/

Abstract

Recent years have witnessed an unprecedented interest in the analysis and generation of images and image sequences using Deep Learning methodologies. In this talk we will focus on some of our recent works in generative models that are primarily aimed at controllable generation. We will first present unsupervised methods for learning non-linear paths in the latent spaces of Generative Adversarial Networks such that following different paths lead to different types of changes (e.g., removing the background, changing head poses, or facial expressions) in the resulting images. Subsequently, we will present a method that allows local editing by finding a Parts and Appearances decomposition in the GAN latent space. Then, we will present recent works in which supervision comes from language models, such that the generation can be controlled by Natural Language sentences. Finally, we will touch on potential applications.