Generative Models for Visual Creation and Editing

Abstract:

Recent advancements in generative models have shown monumental progress ‑ enabling arbitrary generation of text, images, and a handful of other modalities. This begs the question: what's next? In this talk, I'll discuss the state of the art in generative imagery, the significance of generative models in the artistic creative process, as well as recent research in making these models more useful for visual tasks like creation, editing, and analysis.

Bio:

Aleksander Hołyński is a research scientist at Google Research and a postdoctoral scholar at UC Berkeley, working with Alexei Efros and Angjoo Kanazawa. Previously, he was PhD student at the University of Washington, advised by Steve Seitz, Brian Curless, and Rick Szeliski. His research spans computer graphics, computer vision, and machine learning, with a focus on 3D reconstruction, motion estimation, image-based rendering, and generative models. https://holynski.org/