Personal AI, on device.
Exploring how smartphone sensing can make language models more context-aware.
CONTEXT-AWARE INTELLIGENCE
Sensing→
03 / HUMAN-CENTRED COMPUTINGOn-device
LLM
→PersonalisationLLM
Context makes a difference.
This research demo combines on-device language models with smartphone sensing to explore context-aware, personalised services.
Keeping intelligence close.
The framework brings together sensor data and prompt engineering on a smartphone. A university-student case study demonstrates tailored recommendations, while the research examines trade-offs around privacy, latency, cost and energy consumption.
THE IDEA
Personal context + local intelligence → More relevant assistance
A starting point for further research.
This is a research demonstration. Broader sensor inputs and more extensive user studies are identified as future work.
Read the research.
Enabling On-Device LLMs Personalization with Smartphone Sensing.
Shiquan Zhang and collaborators · 2024.