Power efficiency

Power efficiency is essential for AI. We research deep learning techniques for power-efficient implementations and optimization across hardware, algorithms, and software—with a focus on fundamentals to accelerate deep learning workloads at low power to maximize overall efficiency.

Customization & Personalization

Through continuous learning, AI brings personalization to our devices, which is essential for superior user experiences. We research continuous learning techniques that customize user behavior and preferences, bringing a human-like understanding and behavior to IoT devices. 

Scalability

To make AI ubiquitous, we focus on making efficient hardware, algorithmic advancements, and software tools available to developers and OEMs. Voice recognition, environmental perception, energy management...AI is expanding into more areas based on our complete solution.

Security & Privacy

From enterprise to residential applications, IoT connects devices, but AI at the edge is what makes them smart. Our AI Engine enables better security and privacy to IoT devices combined with faster processing and a reduced need for network bandwidth.