Attention-Based Feature Compression for CNN Inference Offloading in Edge Computing

TY - GEN

T1 - Attention-Based Feature Compression for CNN Inference Offloading in Edge Computing

AU - Li, Nan

AU - Iosifidis, Alexandros

AU - Zhang, Qi

PY - 2023/10

Y1 - 2023/10

N2 - This paper studies the computational offloading of CNN inference in device-edge co-inference systems. Inspired by the emerging paradigm semantic communication, we propose a novel autoencoder-based CNN architecture (AECNN), for effective feature extraction at end-device. We design a feature compression module based on the channel attention method in CNN, to compress the intermediate data by selecting the most important features. To further reduce communication overhead, we can use entropy encoding to remove the statistical redundancy in the compressed data. At the receiver, we design a lightweight decoder to reconstruct the intermediate data through learning from the received compressed data to improve accuracy. To fasten the convergence, we use a step-by-step approach to train the neural networks obtained based on ResNet-50 architecture. Experimental results show that AECNN can compress the intermediate data by more than 256 × with only about 4% accuracy loss, which outperforms the state-of-the-art work, BottleNet++. Compared to offloading inference task directly to edge server, AECNN can complete inference task earlier, in particular, under poor wireless channel condition, which highlights the effectiveness of AECNN in guaranteeing higher accuracy within time constraint.

AB - This paper studies the computational offloading of CNN inference in device-edge co-inference systems. Inspired by the emerging paradigm semantic communication, we propose a novel autoencoder-based CNN architecture (AECNN), for effective feature extraction at end-device. We design a feature compression module based on the channel attention method in CNN, to compress the intermediate data by selecting the most important features. To further reduce communication overhead, we can use entropy encoding to remove the statistical redundancy in the compressed data. At the receiver, we design a lightweight decoder to reconstruct the intermediate data through learning from the received compressed data to improve accuracy. To fasten the convergence, we use a step-by-step approach to train the neural networks obtained based on ResNet-50 architecture. Experimental results show that AECNN can compress the intermediate data by more than 256 × with only about 4% accuracy loss, which outperforms the state-of-the-art work, BottleNet++. Compared to offloading inference task directly to edge server, AECNN can complete inference task earlier, in particular, under poor wireless channel condition, which highlights the effectiveness of AECNN in guaranteeing higher accuracy within time constraint.

KW - CNN inference

KW - Edge computing

KW - feature compression

KW - semantic communication

KW - service reliability

UR - http://www.scopus.com/inward/record.url?scp=85177841558&partnerID=8YFLogxK

U2 - 10.1109/ICC45041.2023.10279110

DO - 10.1109/ICC45041.2023.10279110

M3 - Article in proceedings

T3 - I E E E International Conference on Communications

SP - 967

EP - 972

BT - ICC 2023 - IEEE International Conference on Communications

A2 - Zorzi, Michele

A2 - Tao, Meixia

A2 - Saad, Walid

PB - IEEE

T2 - IEEE International Conference on Communications

Y2 - 28 May 2023 through 1 June 2023

ER -