Title: Synthetic and Fake Worlds from a Generative Perspective Abstract: Generative approaches enable creating numerous scenarios coherent with the reality, as long as the representations are good approximations of the real world. In the first part of this talk, I will discuss some of my work in extracting generative representations from 2D and 3D data for digitizing the world for mapping, modeling, and reconstruction, of spatial data and urban models; combining computer vision, machine learning, and computational geometry for shape understanding. Even though such generative approaches are positively impacting some industries like entertainment, they are also used with malicious intent, such as deep fakes. In the second part of the talk, we will shift gears and enter "defense against dark arts" class, introducing our initiatives in generating, detecting, and tracking deep fakes in the wild with high accuracy. Transitioning to 3D, the talk will conclude with a showcase of our AR/VR productions from the world’s largest volumetric capture stage at Intel Studios, as an example of real-world use cases of such AI approaches.

Be the first to comment