Print Email Facebook Twitter ElasticDNN Title ElasticDNN: On-Device Neural Network Remodeling for Adapting Evolving Vision Domains at Edge Author Zhang, Qinglong (Beijing Institute of Technology) Han, Rui (Beijing Institute of Technology) Liu, Chi Harold (Beijing Institute of Technology) Wang, Guoren (Beijing Institute of Technology) Chen, Lydia Y. (TU Delft Data-Intensive Systems) Date 2024 Abstract Executing deep neural networks (DNN) based vision tasks on edge devices encounters challenging scenarios of significant and continually evolving data domains (e.g. background or subpopulation shift). With limited resources, the state-of-the-art domain adaptation (DA) methods either cause high training overheads on large DNN models, or incur significant accuracy losses when adapting small/compressed models in an online fashion. The inefficient resource scheduling among multiple applications further degrades their overall model accuracy. In this paper, we present ElasticDNN, a framework that enables online DNN remodeling for applications encountering evolving domain drifts at edge. Its first key component is the master-surrogate DNN models, which can dynamically generate a small surrogate DNN by retaining and training the large master DNN's most relevant regions pertinent to the new domain. The second novelty of ElasticDNN is the filter-grained resource scheduling, which allocates GPU resources based on online accuracy estimation and DNN remodeling of co-running applications. We fully implement ElasticDNN and demonstrate its effectiveness through extensive experiments. The results show that, compared to existing online DA methods using the same model sizes, ElasticDNN improves accuracy by 23.31% and reduces adaption time by 35.67x. In the more challenging multi-application scenario, ElasticDNN improves accuracy by an average of 25.91%. Subject deep neural networksdomain adaptationedge vision To reference this document use: http://resolver.tudelft.nl/uuid:b7da61f1-ee07-40eb-89f5-f6532d9353a7 DOI https://doi.org/10.1109/TC.2024.3375608 Embargo date 2024-09-16 ISSN 0018-9340 Source IEEE Transactions on Computers, 73 (6), 1616-1630 Bibliographical note Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public. Part of collection Institutional Repository Document type journal article Rights © 2024 Qinglong Zhang, Rui Han, Chi Harold Liu, Guoren Wang, Lydia Y. Chen Files file embargo until 2024-09-16