Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Cao, Bryan Bo, Sharma, Abhinav, Singh, Manavjeet, Gandhi, Anshul, Das, Samir, Jain, Shubham
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2410.11233
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912073274884096
author Cao, Bryan Bo
Sharma, Abhinav
Singh, Manavjeet
Gandhi, Anshul
Das, Samir
Jain, Shubham
author_facet Cao, Bryan Bo
Sharma, Abhinav
Singh, Manavjeet
Gandhi, Anshul
Das, Samir
Jain, Shubham
contents Edge computing has emerged as an alternative to reduce transmission and processing delay and preserve privacy of the video streams. However, the ever-increasing complexity of Deep Neural Networks (DNNs) used in video-based applications (e.g. object detection) exerts pressure on memory-constrained edge devices. Model merging is proposed to reduce the DNNs' memory footprint by keeping only one copy of merged layers' weights in memory. In existing model merging techniques, (i) only architecturally identical layers can be shared; (ii) requires computationally expensive retraining in the cloud; (iii) assumes the availability of ground truth for retraining. The re-evaluation of a merged model's performance, however, requires a validation dataset with ground truth, typically runs at the cloud. Common metrics to guide the selection of shared layers include the size or computational cost of shared layers or representation size. We propose a new model merging scheme by sharing representations (i.e., outputs of layers) at the edge, guided by representation similarity S. We show that S is extremely highly correlated with merged model's accuracy with Pearson Correlation Coefficient |r| > 0.94 than other metrics, demonstrating that representation similarity can serve as a strong validation accuracy indicator without ground truth. We present our preliminary results of the newly proposed model merging scheme with identified challenges, demonstrating a promising research future direction.
format Preprint
id arxiv_https___arxiv_org_abs_2410_11233
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Representation Similarity: A Better Guidance of DNN Layer Sharing for Edge Computing without Training
Cao, Bryan Bo
Sharma, Abhinav
Singh, Manavjeet
Gandhi, Anshul
Das, Samir
Jain, Shubham
Computer Vision and Pattern Recognition
Distributed, Parallel, and Cluster Computing
68M14
C.2.4; I.4.0; I.4.9
Edge computing has emerged as an alternative to reduce transmission and processing delay and preserve privacy of the video streams. However, the ever-increasing complexity of Deep Neural Networks (DNNs) used in video-based applications (e.g. object detection) exerts pressure on memory-constrained edge devices. Model merging is proposed to reduce the DNNs' memory footprint by keeping only one copy of merged layers' weights in memory. In existing model merging techniques, (i) only architecturally identical layers can be shared; (ii) requires computationally expensive retraining in the cloud; (iii) assumes the availability of ground truth for retraining. The re-evaluation of a merged model's performance, however, requires a validation dataset with ground truth, typically runs at the cloud. Common metrics to guide the selection of shared layers include the size or computational cost of shared layers or representation size. We propose a new model merging scheme by sharing representations (i.e., outputs of layers) at the edge, guided by representation similarity S. We show that S is extremely highly correlated with merged model's accuracy with Pearson Correlation Coefficient |r| > 0.94 than other metrics, demonstrating that representation similarity can serve as a strong validation accuracy indicator without ground truth. We present our preliminary results of the newly proposed model merging scheme with identified challenges, demonstrating a promising research future direction.
title Representation Similarity: A Better Guidance of DNN Layer Sharing for Edge Computing without Training
topic Computer Vision and Pattern Recognition
Distributed, Parallel, and Cluster Computing
68M14
C.2.4; I.4.0; I.4.9
url https://arxiv.org/abs/2410.11233