Our research focuses on developing energy-efficient foundational models that can power the next generation of embodied AI systems.
Developing novel network architectures that achieve high performance with minimal computational overhead.
Creating foundational models specifically designed for robots and physical AI systems.
Innovating training techniques that reduce energy consumption without sacrificing model quality.
Optimizing models for deployment on resource-constrained edge devices.
J. Ortiz, M. Chen, E. Rodriguez
In Preparation
We present our research into novel architectures for embodied AI designed to achieve competitive performance while targeting significant reductions in computational requirements.
M. Chen, E. Rodriguez, J. Ortiz
Working Paper
This paper explores new training methodologies designed to reduce the carbon footprint of developing large AI models.
E. Rodriguez, J. Ortiz
Position Paper
A comprehensive framework proposal for developing embodied AI systems with minimal carbon emissions throughout their lifecycle.
We believe in sharing our research openly with the community. All our publications are freely accessible, and we actively contribute to open-source initiatives.