Weight-sharing neural architecture search: A battle to shrink the optimization gap

L Xie, X Chen, K Bi, L Wei, Y Xu, L Wang… - ACM Computing …, 2021 - dl.acm.org
Neural architecture search (NAS) has attracted increasing attention. In recent years,
individual search methods have been replaced by weight-sharing search methods for higher …

Lightweight neural architecture search for temporal convolutional networks at the edge

M Risso, A Burrello, F Conti, L Lamberti… - IEEE Transactions …, 2022 - ieeexplore.ieee.org
Neural Architecture Search (NAS) is quickly becoming the go-to approach to optimize the
structure of Deep Learning (DL) models for complex tasks such as Image Classification or …

Wisdom of committees: An overlooked approach to faster and more accurate models

X Wang, D Kondratyuk, E Christiansen… - arXiv preprint arXiv …, 2020 - arxiv.org
Committee-based models (ensembles or cascades) construct models by combining existing
pre-trained ones. While ensembles and cascades are well-known techniques that were …

End-to-End Neural Network Compression via l1/l2 Regularized Latency Surrogates

A Nasery, H Shah, AS Suggala… - Proceedings of the IEEE …, 2024 - openaccess.thecvf.com
Neural network (NN) compression via techniques such as pruning quantization requires
setting compression hyperparameters (eg number of channels to be pruned bitwidths for …

[PDF][PDF] Multiple networks are more efficient than one: Fast and accurate models via ensembles and cascades

X Wang, D Kondratyuk, KM Kitani… - arXiv preprint arXiv …, 2020 - researchgate.net
Recent work on efficient neural network architectures focuses on discovering a solitary
network that can achieve superior computational efficiency and accuracy. While this …

D-DARTS: Distributed differentiable architecture search

A Heuillet, H Tabia, H Arioui, K Youcef-Toumi - Pattern Recognition Letters, 2023 - Elsevier
Abstract Differentiable ARchiTecture Search (DARTS) is one of the most trending Neural
Architecture Search (NAS) methods. It drastically reduces search cost by resorting to weight …

Art Design Teaching Based on the Multidata Fusion Algorithm and Virtual Simulation Technology

B Liang - Mobile Information Systems, 2022 - Wiley Online Library
Virtual Reality (VR) technologies are widely applied to teaching art design. VR has been
created with high‐level techniques that create the artificial environment to support virtual …

End-to-End Neural Network Compression via Regularized Latency Surrogates

A Nasery, H Shah, AS Suggala, P Jain - arXiv preprint arXiv:2306.05785, 2023 - arxiv.org
Neural network (NN) compression via techniques such as pruning, quantization requires
setting compression hyperparameters (eg, number of channels to be pruned, bitwidths for …

Optimizing AI at the Edge: from network topology design to MCU deployment

A Burrello - 2023 - amsdottorato.unibo.it
The first topic analyzed in the thesis will be Neural Architecture Search (NAS). I will focus on
two different tools that I developed, one to optimize the architecture of Temporal …

[PDF][PDF] Ottimizzazione automatizzata dei modelli DNN per il conteggio delle persone con sensori a infrarossi tramite NAS= Automated Optimization of DNN Models for …

S Mollaei - 2023 - webthesis.biblio.polito.it
The growing popularity of Neural Architecture Search (NAS) is changing optimization
strategies in Deep Learning (DL). While NAS has typically been utilized for tasks such as …