top of page

synthetic intelligence

for learning at the edge



On-device learning at the edge  

  • Demands unsustainable levels of compute resources for ML and AI training in the cloud

  • Requires collection, transfer, and storage of massive datasets

  • Is limited by data latency, which precludes real-time decision making and inflates cost while increasing privacy and security risks



Synthetic intelligence

  • A novel, Goldilocks paradigm combining the robustness and speed of hardware with the low-power, adaptability and complex architectures of wetware 

  • Energy-efficient information processing in scalable and reconfigurable functional materials enabling real-time learning to edge data streams

  • Co-designed hardware and software solutions to push the limits of classical computation beyond current analog and neuromorphic hardware

bottom of page