Lightning Talk: Efficient Embedded Machine Learning Deployment on Edge and IoT Devices

There has been rapid growth in the use of machine learning (ML) software in emerging edge and IoT systems. ML software deployments enable analytics and pattern recognition for multi-modal data (e.g., audio, images/video, wireless signals, air quality) obtained from embedded sensors and transceivers....

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2023 60th ACM/IEEE Design Automation Conference (DAC) s. 1 - 2
Hlavní autor: Pasricha, Sudeep
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 09.07.2023
Témata:
edge computing
embedded software
Image edge detection
Indoor navigation
IoT computing
Machine learning
model optimizations
Software
Transceivers
Wireless communication
Wireless sensor networks
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:There has been rapid growth in the use of machine learning (ML) software in emerging edge and IoT systems. ML software deployments enable analytics and pattern recognition for multi-modal data (e.g., audio, images/video, wireless signals, air quality) obtained from embedded sensors and transceivers. However, resource constraints in edge and IoT platforms make it challenging to meet quality-of-service and real-time goals. The growing complexity of ML also exacerbates these issues. We discuss the challenges of ML software deployment in edge and IoT platforms, present strategies to ease deployment, and discuss case studies from the automotive, indoor navigation, and hardware/software co-design domains.
DOI:10.1109/DAC56929.2023.10247845