Print Email Facebook Twitter Characterising the Role of Pre-Processing Parameters in Audio-based Embedded Machine Learning Title Characterising the Role of Pre-Processing Parameters in Audio-based Embedded Machine Learning Author Hutiri, Wiebke (TU Delft Information and Communication Technology) Mathur, Akhil (Nokia Bell Labs) Ding, Aaron Yi (TU Delft Information and Communication Technology) Kawsar, F. (Nokia Bell Labs) Date 2021 Abstract When deploying machine learning (ML) models on embedded and IoT devices, performance encompasses more than an accuracy metric: inference latency, energy consumption, and model fairness are necessary to ensure reliable performance under heterogeneous and resource-constrained operating conditions. To this end, prior research has studied model-centric approaches, such as tuning the hyperparameters of the model during training and later applying model compression techniques to tailor the model to the resource needs of an embedded device. In this paper, we take a data-centric view of embedded ML and study the role that pre-processing parameters in the data pipeline can play in balancing the various performance metrics of an embedded ML system. Through an in-depth case study with audio-based keyword spotting (KWS) models, we show that pre-processing parameter tuning is a remarkable tool that model developers can adopt to trade-off between a model's accuracy, fairness, and system efficiency, as well as to make an embedded ML model resilient to unseen deployment conditions. Subject audio keyword spottingembedded machine learningfairnesspre-processing parameters To reference this document use: http://resolver.tudelft.nl/uuid:5d18bdbd-ae5f-47f5-961a-85506008a815 DOI https://doi.org/10.1145/3485730.3493448 Publisher Association for Computing Machinery (ACM) ISBN 9781450390972 Source SenSys 2021 - Proceedings of the 2021 19th ACM Conference on Embedded Networked Sensor Systems Event 19th ACM Conference on Embedded Networked Sensor Systems, SenSys 2021, 2021-11-15 → 2021-11-17, Coimbra, Portugal Series SenSys 2021 - Proceedings of the 2021 19th ACM Conference on Embedded Networked Sensor Systems Part of collection Institutional Repository Document type conference paper Rights © 2021 Wiebke Hutiri, Akhil Mathur, Aaron Yi Ding, F. Kawsar Files PDF 3485730.3493448.pdf 788.57 KB Close viewer /islandora/object/uuid:5d18bdbd-ae5f-47f5-961a-85506008a815/datastream/OBJ/view