AdaShadow: Responsive Test-time Model Adaptation in Non-stationary Mobile Environments

Guardat en:
Dades bibliogràfiques
Publicat a:arXiv.org (Oct 10, 2024), p. n/a
Autor principal: Cheng, Fang
Altres autors: Liu, Sicong, Zhou, Zimu, Guo, Bin, Tang, Jiaqi, Ma, Ke, Yu, Zhiwen
Publicat:
Cornell University Library, arXiv.org
Matèries:
Accés en línia:Citation/Abstract
Full text outside of ProQuest
Etiquetes: Afegir etiqueta
Sense etiquetes, Sigues el primer a etiquetar aquest registre!

MARC

LEADER 00000nab a2200000uu 4500
001 3116452596
003 UK-CbPIL
022 |a 2331-8422 
024 7 |a 10.1145/3666025.3699339  |2 doi 
035 |a 3116452596 
045 0 |b d20241010 
100 1 |a Cheng, Fang 
245 1 |a AdaShadow: Responsive Test-time Model Adaptation in Non-stationary Mobile Environments 
260 |b Cornell University Library, arXiv.org  |c Oct 10, 2024 
513 |a Working Paper 
520 3 |a On-device adapting to continual, unpredictable domain shifts is essential for mobile applications like autonomous driving and augmented reality to deliver seamless user experiences in evolving environments. Test-time adaptation (TTA) emerges as a promising solution by tuning model parameters with unlabeled live data immediately before prediction. However, TTA's unique forward-backward-reforward pipeline notably increases the latency over standard inference, undermining the responsiveness in time-sensitive mobile applications. This paper presents AdaShadow, a responsive test-time adaptation framework for non-stationary mobile data distribution and resource dynamics via selective updates of adaptation-critical layers. Although the tactic is recognized in generic on-device training, TTA's unsupervised and online context presents unique challenges in estimating layer importance and latency, as well as scheduling the optimal layer update plan. AdaShadow addresses these challenges with a backpropagation-free assessor to rapidly identify critical layers, a unit-based runtime predictor to account for resource dynamics in latency estimation, and an online scheduler for prompt layer update planning. Also, AdaShadow incorporates a memory I/O-aware computation reuse scheme to further reduce latency in the reforward pass. Results show that AdaShadow achieves the best accuracy-latency balance under continual shifts. At low memory and energy costs, Adashadow provides a 2x to 3.5x speedup (ms-level) over state-of-the-art TTA methods with comparable accuracy and a 14.8% to 25.4% accuracy boost over efficient supervised methods with similar latency. 
653 |a Accuracy 
653 |a Augmented reality 
653 |a Testing time 
653 |a Applications programs 
653 |a Energy costs 
653 |a Parameter sensitivity 
653 |a Mobile computing 
653 |a Adaptation 
653 |a Resource scheduling 
653 |a User experience 
653 |a Computer aided scheduling 
653 |a Estimation 
653 |a Back propagation 
653 |a Uniqueness 
653 |a Run time (computers) 
700 1 |a Liu, Sicong 
700 1 |a Zhou, Zimu 
700 1 |a Guo, Bin 
700 1 |a Tang, Jiaqi 
700 1 |a Ma, Ke 
700 1 |a Yu, Zhiwen 
773 0 |t arXiv.org  |g (Oct 10, 2024), p. n/a 
786 0 |d ProQuest  |t Engineering Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3116452596/abstract/embedded/L8HZQI7Z43R0LA5T?source=fedsrch 
856 4 0 |3 Full text outside of ProQuest  |u http://arxiv.org/abs/2410.08256