DiTMoS: Delving into Diverse Tiny-Model Selection on Microcontrollers
Published in PerCom 2024 (Mark Weiser Best Paper Award), 2024
Recommended citation: Xiao Ma, Shengfeng He, Hezhe Qiao, Dong Ma. (2024). "DiTMoS: Delving into Diverse Tiny-Model Selection on Microcontrollers." PerCom 2024. https://arxiv.org/abs/2403.09035
🏆 Mark Weiser Best Paper Award Winner
This paper introduces DiTMoS, a comprehensive framework for selecting and deploying diverse tiny machine learning models on resource-constrained microcontrollers.
Abstract
The deployment of machine learning models on microcontrollers presents unique challenges due to severe resource constraints. DiTMoS addresses these challenges by providing a systematic approach to model selection and optimization.
Key Contributions
- Comprehensive model selection framework for TinyML
- Performance evaluation across diverse microcontroller platforms
- Real-world deployment case studies
- Open-source implementation and benchmarks
Links
Recommended citation: Xiao Ma, Shengfeng He, Hezhe Qiao, Dong Ma. (2024). “DiTMoS: Delving into Diverse Tiny-Model Selection on Microcontrollers.” PerCom 2024.