GAT TransPruning: progressive channel pruning strategy combining graph attention network and transformer [PDF]
PeerJ Computer ScienceRecently, large-scale artificial intelligence models with billions of parameters have achieved good results in experiments, but their practical deployment on edge computing platforms is often subject to many constraints because of their resource ...
Yu-Chen Lin +2 more
doaj +3 more sources
Polarization-Regularized Adversarial Pruning for Efficient Radio Frequency Fingerprint Identification on IoT Devices [PDF]
SensorsRadio frequency fingerprint identification (RFFI) based on physical-layer characteristics provides a reliable solution for secure authentication of Internet of Things (IoT) devices.
Caidan Zhao +4 more
doaj +2 more sources
An Accelerated MMP With a Pruning Tree Strategy [PDF]
IEEE Access, 2019 Multipath matching pursuit (MMP) has been developed to solve the sparse signal recovery problems in compressed sensing, which can generate recovery error less than the traditional orthogonal matching pursuit type algorithms in terms of mean square error.
Yuli Fu +4 more
doaj +2 more sources
Is it possible to prevent excessive synaptic pruning in schizophrenia? Possibilities and limitations [PDF]
Frontiers in Synaptic NeuroscienceBackgroundSynaptic pruning is a critical neurodevelopmental process that eliminates redundant or weak synaptic connections to optimize brain circuitry. In schizophrenia, converging evidence from imaging, genetic, and postmortem studies suggests that this
Agnieszka Pawlak +7 more
doaj +2 more sources
Mechanical Pruning of ‘Clemenules’ Mandarins in Spain: Yield Effects and Economic Analysis
Agronomy, 2022 Pruning is one of the most expensive tasks in citrus production, and its mechanization could increase the productivity and competitiveness of citrus farms.
Alberto Fonte +4 more
doaj +1 more source
An Efficient Method for Mining Periodic Cliques [PDF]
Jisuanji gongcheng, 2023 Periodic cliques are complete subgraphs that meet specific periodic requirements when they appear on the temporal network.Periodic cliques mining is used to mine periodic cliques in the temporal graph.Considering the low efficiency of the existing ...
DU Ming, HAO Yan, ZHOU Junfeng, TAN Yuting
doaj +1 more source
Pruning Deep Convolutional Neural Networks Architectures with Evolution Strategy [PDF]
Information Sciences, 2021 Currently, Deep Convolutional Neural Networks (DCNNs) are used to solve all kinds of problems in the field of machine learning and artificial intelligence due to their learning and adaptation capabilities. However, most successful DCNN models have a high computational complexity making them difficult to deploy on mobile or embedded platforms.
Fernandes, Francisco E. jun. +1 more
openaire +2 more sources
Incomplete Information Game Algorithm Based on Expectimax Search and Double DQN [PDF]
Jisuanji gongcheng, 2021 As a typical incomplete information game, mahjong is mainly realized by the traditional Expectimax search algorithm, whose pruning strategy and valuation function design based on artificial prior knowledge and thus cause unreasonable assumptions and ...
LEI Jiewei, WANG Jiayang, REN Hang, YAN Tianwei, HUANG Wei
doaj +1 more source
Pre-indexing Pruning Strategies [PDF]
, 2020 We explore different techniques for pruning an inverted index in advance, that is, without building the full index. These techniques provide interesting trade-offs between index size, answer quality and query coverage. We experimentally analyze them in a large public web collection with two different query logs.
Soner Altin +2 more
openaire +1 more source
Towards Robust Pruning: An Adaptive Knowledge-Retention Pruning Strategy for Language Models
Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, 2023 The pruning objective has recently extended beyond accuracy and sparsity to robustness in language models. Despite this, existing methods struggle to enhance robustness against adversarial attacks when continually increasing model sparsity and require a retraining process.
Li, Jianwei +3 more
openaire +2 more sources