Ziqi Ma

@ziqi__ma

PhD student @caltech, research intern @theworldlabs, previously @AIatMeta, @microsoft, @uchicago. https://ziqi-ma.github.io

following
followers

LoGo aims to improve 3D consistency in long AI-generated videos

A researcher describes a spatially localized 3D reward model designed to show where generated scenes go wrong.

5 Sources, 14d ago, updated 14d ago

DynaTokens is headed to NeurIPS

A post announcing DynaTokens says its approach trains a few tokens at test time to enable scene dynamics in world models while preserving camera control.

3 Sources, 21d ago, first seen 21d ago

World Labs Unveils Atlas World Model

Atlas generates videos with exact camera control and reconstructs 3D scenes from input images.

126 Sources, 49d ago, first seen 49d ago

Featured in

LoGo aims to improve 3D consistency in long AI-generated videos

DynaTokens is headed to NeurIPS

World Labs Unveils Atlas World Model

Featured in

LoGo aims to improve 3D consistency in long AI-generated videos

DynaTokens is headed to NeurIPS

World Labs Unveils Atlas World Model