Hierarchical Inference (HI) has emerged as a promising approach for efficient distributed inference between end devices deployed with small pre-trained Deep Learning models and edge/cloud servers running large DL models. Under HI, a device uses the local DL model to perform inference on the data samples it collects, and only the data samples on which this inference is likely to be incorrect are offloaded to a remote DL model running on the server. Thus, gauging the likelihood of incorrect local inference is key to implementing HI. A natural approach is to compute a confidence metric for the local DL inference and then use a threshold on this confidence metric to determine whether to offload or not. Recently, the HI online learning problem was studied to learn an optimal threshold for the confidence metric over a sequence of data samples collected over time. However, existing algorithms have computation complexity that grows with the number of rounds and do not exhibit a sub-linear regret bound. In this work, we propose two algorithms, Hedge-HI and Hedge-HI-Restart, and prove that they achieve sublinear regret bounds. The latter, which is novel, also enjoys significantly lower computational complexity that grows sublinearly with the number of rounds. Using runtime measurements on Raspberry Pi, we demonstrate that Hedge-HI-Restart has a runtime lower by order of magnitude and achieves cumulative loss close to that of the alternatives.

About Ghina Al Atat

Ghina is a PhD researcher in the Opportunistic Architectures Lab at IMDEA Networks Institute in Madrid, Spain, and a PhD student in the Department of Telematics Engineering at UC3M, Madrid. She holds a bachelor’s degree in Physics and a master’s degree in Computational Science from the American University of Beirut, Lebanon. Her current research focuses on distributed machine learning inference to support ML applications in the IoT and edge computing ecosystem, with a particular interest in edge intelligence.

Este evento se impartirá en inglés