CVPR Daily - Thursday

3.1.1: Diffusion Autoencoders: Toward a Meaningful and Decodable Representation 3.1.2: Reversible Vision Transformers 3.2.2: StyleSDF: High-Resolution 3D-Consistent Image and Geometry Generation 3.1: Panoptic Neural Fields: A Semantic Object-Aware Neural Scene Representation 3.1: GIFS: Neural Implicit Function for General Shape Representation 3.2: Primitive3D: 3D Object Dataset Synthesis From Randomly Assembled Primitives 3.2: Self-Supervised Neural Articulated Shape and Appearance Models Anh’s picks of the day: Ành Thai is currently a PhD student at Georgia Institute of Technology, USA, advised by Prof. James Rehg. Her research generally focuses on object representation learning inspired by developmental psychology. “ Motivated by the question: “Can we devise artificial agents that can learn and accumulate knowledge in the same manner as infants?”, my research works have investigated object representation learning in settings that more closely resemble the way young children learn such as continual, self-supervised, and few- shot learning. For today, Thursday 23 2 Anh’s Picks DAILY CVPR Thursday I’m excited that we are finally back to attending conferences in person. This is my second in- person CVPR and I am looking forward to meeting everyone after such a long time. CVPR 2019 was my first conference ever and having the chance to be engaged with the computer vision community has motivated me to do research every day. I love playing musical instruments and analog photography. I also like traveling and kayaking. ” Orals Posters