Katerina Fragkiadaki

Katerina Fragkiadaki is the JPMorgan Chase Associate Professor of Computer Science in the Machine Learning Department at Carnegie Mellon University. She works at the intersection of Artificial Intelligence, Computer Vision, Machine Learning, Language Understanding, and Robotics.

• PhD from GRASP Lab, University of Pennsylvania
• Postdoctoral researcher at UC Berkeley (with Jitendra Malik) and Google Research
• Recipient of NSF CAREER, DARPA Young Investigator, Amazon, Google, Sony, UPMC, and AFOSR awards
• Organizer of CoRL 2023 Workshop on Generalist Robots
• ICLR 2024 Program Chair, multiple area chair roles

Her research group focuses on developing machines that autonomously improve world models through human-environment interactions, with specific emphasis on:

• Representation learning and video understanding
• 2D/3D unified vision-language models
• Generative simulation and reinforcement learning
• Real2Sim/Sim2Real robot learning
• Continual learning and spatial common sense
• 3D scene reconstruction and dynamics

Recent publications highlight advancements in:

• 3D mesh generation with compositional transformers
• Unified 2D/3D perception frameworks
• Physics-aware generative models
• Diffusion-based robotic manipulation policies
• Embodied agents with memory prompting

Awards include:

• 2024: DARPA Young Investigator Award
• 2023: Amazon Faculty Award
• 2022: Sony Faculty Research Award
• 2021: UPMC Faculty Research Award
• 2020: NSF CAREER Award
• 2019: Google Faculty Award

Key collaborations span institutions including UC Berkeley, Google Research, Stanford, MIT, and University of Tsukuba. Her work bridges theoretical innovation with practical applications in:

• Autonomous robot manipulation
• 4D world modeling
• Language-grounded perception
• Visual dynamics prediction
• Embodied program synthesis
• Physics-based simulation engines