Improved Decision Module Selection for Hierarchical Inference in Resource-Constrained Edge Devices

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Behera, Adarsh Prasad, Morabito, Roberto, Widmer, Joerg, Champati, Jaya Prakash
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929385392570368
author Behera, Adarsh Prasad
Morabito, Roberto
Widmer, Joerg
Champati, Jaya Prakash
author_facet Behera, Adarsh Prasad
Morabito, Roberto
Widmer, Joerg
Champati, Jaya Prakash
contents The Hierarchical Inference (HI) paradigm employs a tiered processing: the inference from simple data samples are accepted at the end device, while complex data samples are offloaded to the central servers. HI has recently emerged as an effective method for balancing inference accuracy, data processing, transmission throughput, and offloading cost. This approach proves particularly efficient in scenarios involving resource-constrained edge devices, such as IoT sensors and micro controller units (MCUs), tasked with executing tinyML inference. Notably, it outperforms strategies such as local inference execution, inference offloading to edge servers or cloud facilities, and split inference (i.e., inference execution distributed between two endpoints). Building upon the HI paradigm, this work explores different techniques aimed at further optimizing inference task execution. We propose and discuss three distinct HI approaches and evaluate their utility for image classification.
format Preprint
id arxiv_https___arxiv_org_abs_2406_09424
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Improved Decision Module Selection for Hierarchical Inference in Resource-Constrained Edge Devices
Behera, Adarsh Prasad
Morabito, Roberto
Widmer, Joerg
Champati, Jaya Prakash
Distributed, Parallel, and Cluster Computing
The Hierarchical Inference (HI) paradigm employs a tiered processing: the inference from simple data samples are accepted at the end device, while complex data samples are offloaded to the central servers. HI has recently emerged as an effective method for balancing inference accuracy, data processing, transmission throughput, and offloading cost. This approach proves particularly efficient in scenarios involving resource-constrained edge devices, such as IoT sensors and micro controller units (MCUs), tasked with executing tinyML inference. Notably, it outperforms strategies such as local inference execution, inference offloading to edge servers or cloud facilities, and split inference (i.e., inference execution distributed between two endpoints). Building upon the HI paradigm, this work explores different techniques aimed at further optimizing inference task execution. We propose and discuss three distinct HI approaches and evaluate their utility for image classification.
title Improved Decision Module Selection for Hierarchical Inference in Resource-Constrained Edge Devices
topic Distributed, Parallel, and Cluster Computing
url https://arxiv.org/abs/2406.09424