Hsieh, Cheng YenCheng YenHsiehYU-CHUAN CHUANGAN-YEU(ANDY) WU2023-07-172023-07-172022-01-01978166548547021610363https://scholars.lib.ntu.edu.tw/handle/123456789/633754Most existing studies improve the efficiency of Split learning (SL) by compressing the transmitted features. However, most works focus on dimension-wise compression that transforms high-dimensional features into a low-dimensional space. In this paper, we propose circular convolution-based batch-wise compression for SL (C3-SL) to compress multiple features into one single feature. To avoid information loss while merging multiple features, we exploit the quasi-orthogonality of features in high-dimensional space with circular convolution and superposition. To the best of our knowledge, we are the first to explore the potential of batch-wise compression under the SL scenario. Based on the simulation results on CIFAR-10 and CIFAR-100, our method achieves a 16x compression ratio with negligible accuracy drops compared with the vanilla SL. Moreover, C3-SL significantly reduces 1152x memory and 2.25x computation overhead compared to the state-of-the-art dimension-wise compression method.Cloud-edge collaborative learning | communication efficiency | data compression | split learning[SDGs]SDG3conference paper10.1109/MLSP55214.2022.99435072-s2.0-85142714860https://api.elsevier.com/content/abstract/scopus_id/85142714860