analiza aiqbench
PrototypeNAS introduces a zero-shot neural architecture search method that automates the optimization of deep learning models for resource-constrained microcontroller units without requiring extensive retraining. By decoupling design from training and utilizing multi-objective optimization, the approach significantly accelerates the deployment of efficient DNNs on edge hardware while maintaining high accuracy. This advancement addresses critical bottlenecks in adapting complex AI models to diverse embedded systems with minimal manual intervention.
PrototypeNAS: Rapid Design of Deep Neural Networks for Microcontroller Units
arXiv:2603.15106v2 Announce Type: cross Abstract: Enabling efficient deep neural network (DNN) inference on edge devices with different hardware constraints is a challenging task that typically requires DNN architectures to be specialized for each device separately. To avoid the huge manual effort, one can use neural architecture search (NAS). However, many existing NAS methods are resource-intensive and time-consuming because they require the training of many different DNNs from scratch. Furthermore, they do not take the resource constraints of the target system into account. To address these shortcomings, we propose PrototypeNAS, a zero-shot NAS method to accelerate and automate the selection, compression, and specialization of DNNs to different target microcontroller units (MCUs). We propose a novel three-step search method that decouples DNN design and specialization from DNN training for a given target platform. First, we present a novel search space that not only cuts out smaller DNNs from a single large architecture, but instead combines the structural optimization of multiple architecture types, as well as optimization of their pruning and quantization configurations. Second, we explore the use of an ensemble of zero-shot proxies during optimization instead of a single one. Third, we propose the use of Hypervolume subset selection to distill DNN architectures from the Pareto front of the multi-objective optimization that represent the most meaningful tradeoffs between accuracy and FLOPs. We evaluate the effectiveness of PrototypeNAS on 12 different datasets in three different tasks: image classification, time series classification, and object detection. Our results demonstrate that PrototypeNAS is able to identify DNN models within minutes that are small enough to be deployed on off-the-shelf MCUs and still achieve accuracies comparable to the performance of large DNN models.
Abstract:Enabling efficient deep neural network (DNN) inference on edge devices with different hardware constraints is a challenging task that typically requires DNN architectures to be specialized for each device separately. To avoid the huge manual effort, one can use neural architecture search (NAS). However, many existing NAS methods are resource-intensive and time-consuming because they require the training of many different DNNs from scratch. Furthermore, they do not take the resource constraints of the target system into account. To address these shortcomings, we propose PrototypeNAS, a zero-shot NAS method to accelerate and automate the selection, compression, and specialization of DNNs to different target microcontroller units (MCUs). We propose a novel three-step search method that decouples DNN design and specialization from DNN training for a given target platform. First, we present a novel search space that not only cuts out smaller DNNs from a single large architecture, but instead combines the structural optimization of multiple architecture types, as well as optimization of their pruning and quantization configurations. Second, we explore the use of an ensemble of zero-shot proxies during optimization instead of a single one. Third, we propose the use of Hypervolume subset selection to distill DNN architectures from the Pareto front of the multi-objective optimization that represent the most meaningful tradeoffs between accuracy and FLOPs. We evaluate the effectiveness of PrototypeNAS on 12 different datasets in three different tasks: image classification, time series classification, and object detection. Our results demonstrate that PrototypeNAS is able to identify DNN models within minutes that are small enough to be deployed on off-the-shelf MCUs and still achieve accuracies comparable to the performance of large DNN models.
Accepted at ECML-PKDD 2026. 18 pages, 7 figures, 4 tables. This work was funded by the European Commission as part of the MANOLO project under the Horizon Europe programme Grant Agreement No.101135782
Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
arXiv:2603.15106 [cs.AI]
(or arXiv:2603.15106v2 [cs.AI] for this version)
Text integral sub licența deschisă a publicației · 323 cuvinte
Subiecte conexe
News
Quantum Computing Report · 25 iul. 2026
Master AI Chip Principles With New IEEE Design Program
IEEE Spectrum · 9 oct. 2026
News
Quantum Computing Report · 7 iun. 2026
The hard physics and complex economics of AI’s insatiable hunger for power
SiliconANGLE · 11 sept. 2026