He talks about Imagen for text-to-image synthesis. Imagen has shown an unprecedented ability to represent and reflect text descriptions in high-fidelity images. It builds on the power of large transformer language models in encoding text and hinges on the strength of diffusion models for high-fidelity image generation. I will describe how text-to-image models have helped create an ecosystem of enthusiastic designers, artists, and researchers. He also concludes by discussing some recent work on image editing, 3d, and video generation.
Mohammad Norouzi is a research scientist at Google Brain in Toronto. He is interested in developing simple and efficient machine learning algorithms that help solve challenging problems across natural language processing, computer vision, and creativity applications. He was a co-developer of Google’s neural machine translation system, SimCLR, and Imagen.