Papers
13
Total Citations
318
H-Index
8
About
Sanja Fidler is a prominent AI researcher whose work spans 3D scene understanding, generative modeling, and human-robot interaction. Best known for her contributions to visual computing and simulation, Fidler has helped push the boundaries of how machines perceive, reconstruct, and synthesize complex environments. Her most impactful work includes physics-based human motion synthesis from video (77 citations), which eliminates the costly dependence on motion capture data, and NeuralField-LDM (43 citations), a hierarchical latent diffusion framework capable of generating richly detailed 3D environments — a breakthrough for virtual reality and robotics simulation. Her research into 3D Gaussian ray tracing (54 citations) advances real-time rendering of particle-based radiance fields, while GameGAN explores learning environment simulators directly from gameplay footage. Fidler has also made notable contributions to human-in-the-loop learning, developing systems that allow non-expert users to teach machines through natural language feedback — a vision deeply tied to accessible, household robotics. Her work on automated robot design via neural graph evolution further reflects her interdisciplinary reach across graphics, simulation, and autonomous systems. Across her career, Fidler has established herself as a leading voice in AI-driven 3D content creation and intelligent simulation.
Research Focus
Key Achievements
Top Papers
- 1Physics-based Human Motion Estimation and Synthesis from Videos77 citations · 2021
- 23D Gaussian Ray Tracing: Fast Tracing of Particle Scenes54 citations · 2024
- 3NeuralField-LDM: Scene Generation with Hierarchical Latent Diffusion Models43 citations · 2023
- 4Teaching Machines to Describe Images with Natural Language Feedback38 citations · 2017
- 5Neural Graph Evolution: Towards Efficient Automatic Robot Design33 citations · 2019
- 6Teaching Machines to Describe Images via Natural Language Feedback32 citations · 2017
- 7Learning Hierarchical Representations of Object Categories for Robot Vision12 citations · 2010
- 8Interactive Annotation of 3D Object Geometry Using 2D Scribbles11 citations · 2020
- 9Learning to Simulate Dynamic Environments With GameGAN5 citations · 2020
- 10