Neural network architecture search algorithm guided by hybrid heuristic information

doi:10.11772/j.issn.1001-9081.2025030300

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (2): 395-405.DOI: 10.11772/j.issn.1001-9081.2025030300

• Artificial intelligence • Previous Articles

Neural network architecture search algorithm guided by hybrid heuristic information

Qianlong XIONG¹^,², Jin QIN¹^,²()

^1.State Key Laboratory of Public Big Data （Guizhou University），Guizhou Guiyang 550025，China
^2.College of Computer Science and Technology，Guizhou University，Guizhou Guiyang 550025，China

Received:2025-03-25 Revised:2025-05-14 Accepted:2025-05-16 Online:2025-05-29 Published:2026-02-10
Contact: Jin QIN
About author:XIONG Qianlong， born in 1997， M. S. candidate. His research interests include evolutionary neural network architecture search.
QIN Jin， born in 1978， Ph. D.， associate professor. His research interests include computational intelligence， reinforcement learning. Email:jqin1@gzu.edu.cn
Supported by:
National Natural Science Foundation of China(62162007);Scientific and Technological Project in Guizhou （Qiankehe Talents Program KJZY ［2025］020）

混合启发信息指导神经网络架构搜索算法

熊前龙¹^,², 秦进¹^,²()

^1.公共大数据国家重点实验室（贵州大学），贵阳 550025
^2.贵州大学计算机科学与技术学院，贵阳 550025

通讯作者: 秦进
作者简介:熊前龙（1997—），男，贵州黔西人，硕士研究生，主要研究方向：进化神经网络架构搜索
秦进（1978—），男，贵州黔西人，副教授，博士，CCF会员，主要研究方向：计算智能、强化学习。Email:jqin1@gzu.edu.cn
基金资助:
国家自然科学基金资助项目(62162007);贵州省科技计划项目(黔科合人才KJZY［2025］020)

Abstract

Abstract:

Aiming at Neural network Architecture Search （NAS） tasks， an NAS guided by Hybrid Heuristic Information （GHHI-NAS） algorithm was proposed. Firstly， a heuristic information construction module integrating prior knowledge and local search feedback was designed to generate multi-dimensional dynamic heuristic indicators， and a hybrid update strategy was combined to guide architecture search， thereby solving the problems of global exploration deficiency and local optimal traps caused by unidirectional updates in traditional NAS effectively. Secondly， a Covariance Matrix Adaptation Evolution Strategy （CMA-ES） update framework was adopted， enhanced by a hybrid fitness evaluation function， so that the algorithm was able to escape from small-model traps during early stage. Finally， a fitness sharing strategy was implemented to realize smooth evaluation of noise and improve population diversity. Additionally， a Monte Carlo swap sampling method with penalty mechanism was introduced to further reduce performance degradation caused by sampling. Experimental results demonstrate that GHHI-NAS has the validation accuracies of 97.55% and 83.44% on the CIFAR-10 and CIFAR-100 datasets， respectively， along with a test error rate of 24.7% on the ImageNet dataset and outstanding performance on the NAS-Bench-201 dataset， which is similar to or slightly surpasses those of Evolutionary NAS （ENAS） algorithm while requiring only 0.12 GPU-Days for search time， achieving low search costs and high test performance.

Key words: Evolutionary Algorithm (EA), Neural network Architecture Search (NAS), hybrid heuristic information, Covariance Matrix Adaptation Evolution Strategy (CMA-ES), Monte Carlo swap sampling

摘要：

针对神经网络架构搜索（NAS）任务，提出一种混合启发信息指导NAS（GHHI-NAS）算法。首先，通过设计融合先验知识与局部搜索反馈的启发信息构造模块，生成多维动态启发指标，并配合混合更新策略指导架构搜索，从而有效解决传统NAS因更新方向单一导致的全局探索不足及局部最优陷阱的问题；其次，使用自适应协方差进化策略（CMA-ES）作为更新框架，并辅以混合适应度评价函数，从而指导算法在初期跳出小模型陷阱；最后，通过适应度共享策略平滑地评价噪声并提升种群多样性。此外，为了进一步降低采样带来的性能损失，提出带惩罚机制的蒙特卡洛交换采样方法。实验结果表明，GHHI-NAS算法在CIFAR-10和CIFAR-100数据集上分别取得了97.55%和83.44%的验证正确率，在ImageNet数据集上取得了24.7%的测试错误率，在NAS-Bench-201数据集上也取得了杰出的表现，接近甚至略优于进化NAS（ENAS）算法，同时搜索时间仅为0.12 GPU-Days，实现了较低的搜索开销和较高水平的测试性能。

关键词: 进化算法, 神经网络架构搜索, 混合启发信息, 自适应协方差策略, 蒙特卡洛交换采样

CLC Number:

TP183

Qianlong XIONG, Jin QIN. Neural network architecture search algorithm guided by hybrid heuristic information[J]. Journal of Computer Applications, 2026, 46(2): 395-405.

熊前龙, 秦进. 混合启发信息指导神经网络架构搜索算法[J]. 《计算机应用》唯一官方网站, 2026, 46(2): 395-405.

Figures/Tables 11

Fig. 1 Schematic diagram of partial channel connection and CFCM fusion

Fig. 2 Flow of proposed algorithm

Tab. 1 Hybrid update formula and its encoding of heuristic information

方法	更新公式	编码
方法1	$α v (1) = α v + p ξ (v)$	000
方法2	$α v + 1 (2) = α v + θ 1 I 1 (v + 1)$	100
方法3	$α v + 1 (3) = α v + θ 2 I 2 (v + 1)$	010
方法4	$α v + 1 (4) = α v + θ 3 I 3 (v)$	001
方法5	$α v + 1 (5) = α v + θ 1 I 1 (v + 1) + θ 2 I 2 (v + 1)$	110
方法6	$α v + 1 (6) = α v + θ 1 I 1 (v + 1) + θ 3 I 3 (v)$	101
方法7	$α v + 1 (7) = α v + θ 2 I 2 (v + 1) + θ 3 I 3 (v)$	011
方法8	$α v + 1 (8) = α v + θ 1 I 1 (v + 1) + θ 2 I 2 (v + 1) + θ 3 I 3 (v)$	111

Tab. 1 Hybrid update formula and its encoding of heuristic information

方法	更新公式	编码
方法1	$α v (1) = α v + p ξ (v)$	000
方法2	$α v + 1 (2) = α v + θ 1 I 1 (v + 1)$	100
方法3	$α v + 1 (3) = α v + θ 2 I 2 (v + 1)$	010
方法4	$α v + 1 (4) = α v + θ 3 I 3 (v)$	001
方法5	$α v + 1 (5) = α v + θ 1 I 1 (v + 1) + θ 2 I 2 (v + 1)$	110
方法6	$α v + 1 (6) = α v + θ 1 I 1 (v + 1) + θ 3 I 3 (v)$	101
方法7	$α v + 1 (7) = α v + θ 2 I 2 (v + 1) + θ 3 I 3 (v)$	011
方法8	$α v + 1 (8) = α v + θ 1 I 1 (v + 1) + θ 2 I 2 (v + 1) + θ 3 I 3 (v)$	111

Tab. 2 Performance comparison of different algorithms on CIFAR dataset

对比算法	搜索开销/GPU-Days	CIFAR-10		CIFAR-100		类型
对比算法	搜索开销/GPU-Days	参数量/10⁶	验证正确率/%	参数量/10⁶	验证正确率/%	类型
NASNet-A^［37］	2 000	3.3	97.35	3.3	83.18	RL
ENAS^［11］	0.5	4.6	97.11	—	—	RL
DARTS（1st）^［12］	0.4	3.4	97.00±0.14	3.4	82.46	GD
DARTS（2nd） ^［12］	1	3.3	97.24±0.09	—	—	GD
ProxylessNAS^［38］	4	5.7	97.92	—	—	GD
PC-DARTS^［16］	0.1	3.6	97.43±0.07	3.6	83.10	GD
P-DARTS^［39］	0.3	3.3±0.21	97.19±0.14	—	—	GD
SNAS^［40］	1.5	2.8	97.15±0.02	2.8	82.45	GD
ADARTS^［41］	0.2	2.9	97.54	2.9	82.94	GD
PA-DARTS^［42］	0.36	3.3	97.42	3.3	83.03	GD
SWD-NAS^［43］	0.13	3.17	97.49	3.17	82.98	GD
DARTS-EAST^［44］	0.59	3.1	97.46±0.07	3.4	83.23	GD
MIG-DARTS^［45］	0.1	3.8	97.56	3.7	83.66	GD
NPENAS^［46］	1.8/6	3.5	97.46±0.10	3.83	84.77	EA
MSNAS^［47］	0.23	3.25	97.32±0.08	3.25	82.80	EA
Evo-OSNet^［48］	0.5	3.3	97.44	3.5	84.16	EA
SI-EvoNAS^［49］	0.458/0.500	1.84	97.31	3.5	84.16	EA
EAEPSO^［50］	2.2	2.94	97.26	2.94	83.06	EA
SPNAS^［51］	1.4/1.6	6.33	98.20	6.74	87.26	EA
GHHI-NAS	0.12	3.77	97.55±0.08	3.83	83.44±0.28	EA

Fig. 3 Output architecture of GHHI-NAS on CIFAR-10 dataset

Tab. 3 Performance comparison of different NAS algorithms on ImageNet dataset

算法	测试误差/%	搜索开销/GPU-Days	参数量/10⁶	类型	算法	测试误差/%	搜索开销/GPU-Days	参数量/10⁶	类型
NASNet-A^［37］	26.0	1 800	5.3	RL	SWD-NAS^［43］	24.5	0.13	6.3	GD
NASNet-B^［37］	27.2	1 800	5.3	RL	DARTS-EAST^［44］	25.0	—	2.3	GD
NASNet-C^［37］	27.5	1 800	4.9	RL	AmoebaNet-A^［18］	26.0	3 150	3.2	EA
DARTS（2nd）^［12］	26.7	1	4.7	GD	EAEPSO^［50］	26.9	4	4.9	EA
SNAS^［40］	27.3	1.5	2.8	GD	SI-EvoNAS^［49］	24.2	0.458	4.7	EA
ProxylessNAS^†［38］	24.9	8.3	7.1	GD	SPNAS^［51］	21.4	1.77	6.6	EA
PC-DARTS^［16］	25.1	0.1	4.7	GD	GHHI-NAS	24.7	0.12	5.3	EA
P-DARTS^［39］	24.4	0.3	5.1	GD	GHHI-NAS^†	25.6	2.4	4.8	EA
PA-DARTS^［42］	24.7	0.4	5.2	GD

Tab.4 Performance comparison of different algorithms on NAS-Bench-201 dataset

方法	CIFAR-10		CIFAR-100		ImageNet16-120		类型
方法	验证正确率	测试正确率	验证正确率	测试正确率	验证正确率	测试正确率	类型
理论最优值	91.61	94.37	73.49	73.51	46.77	47.31
DARTS（1st）^［12］	39.77±0.00	54.30±0.00	15.03±0.00	15.61±0.00	16.43±0.00	16.32±0.00	GD
DARTS（2nd）^［12］	39.77±0.00	54.30±0.00	15.03±0.00	15.61±0.00	16.43±0.00	16.32±0.00	GD
SNAS^［40］	90.10±1.04	92.77±0.83	69.69±2.39	69.34±1.98	42.84±1.79	43.16±2.64	GD
PC-DARTS^［16］	89.96±0.15	93.41±0.30	67.12±0.39	67.48±0.89	40.83±0.08	41.31±0.22	GD
iDARTS^［31］	89.86±0.60	93.58±0.32	70.57±0.24	70.83±0.48	40.38±0.59	40.89±0.68	GD
CyDAS^［52］	91.12±0.44	94.02±0.31	72.12±1.23	71.92±1.30	45.09±0.61	45.51±0.72	GD
NPENAS-CCL^*［46］	91.57±0.13	94.32±0.19	73.48±0.15	73.49±0.23	46.62±0.34	45.61±0.41	GD
DARTS-EAST^［44］	91.44±0.16	94.18±0.25	72.86±0.89	72.97±0.77	46.16±0.29	45.78±0.80	GD
ℓ-DARTS^［53］	91.61±0.00	94.37±0.00	72.75±0.00	73.22±0.00	45.56±0.00	46.71±0.00	GD
FairNAS^［19］	90.07±0.57	93.23±0.18	70.94±0.94	71.00±1.46	41.90±1.00	42.19±0.31	EA
ReNAS^［54］	90.90±0.31	93.99±0.25	71.96±0.99	72.12±0.79	45.85±0.47	45.97±0.49	EA
SAENAS-NE^［55］	91.58±0.09	94.34±0.12	73.46±0.18	73.46±0.20	46.59±0.14	46.36±0.26	EA
SPNAS^［51］	91.57±0.09	94.37±0.00	73.60±0.11	73.75±0.23	46.42±0.32	46.52±0.32	EA
GHHI-NAS	91.44±0.00	94.33±0.00	72.54±0.00	72.89±0.00	46.16±0.00	46.66±0.00	EA

Tab. 5 Performance comparison of different sampling methods

对比方法	搜索阶段		训练阶段
对比方法	验证正确率/%	搜索开销/GPU-Days	验证正确率/%	参数量/10⁶
GHHI-NAS简单采样	85.97	0.09	97.27	4.12
GHHI-NAS交换采样	86.71	0.12	97.55	3.77
DARTS（2nd）^［12］简单采样	87.15	0.57	97.10	2.71
DARTS（2nd）^［12］交换采样	89.75	0.59	97.34	2.80
PC-DARTS^［16］简单采样	85.91	0.09	97.22	3.48
PC-DARTS^［16］交换采样	86.91	0.10	97.41	3.80

Tab. 6 Impact of different hybrid guidance architecture updates of heuristic information on performance of CIFAR-10 dataset

对比方法	验证正确率/%		参数量/10⁶	搜索开销/GPU-Days
对比方法	搜索阶段	训练阶段	参数量/10⁶	搜索开销/GPU-Days
$I 1$	80.51	97.31	4.16	0.062
$I 2$	82.51	96.53	3.48	0.066
$I 3$	78.95	96.98	2.69	0.064
$I 1 + I 2$	84.71	97.31	2.70	0.068
$I 1 + I 3$	81.78	96.54	3.71	0.071
$I 2 + I 3$	84.42	97.07	2.53	0.070
$I 1 + I 2 + I 3$	87.61	97.55	3.77	0.120

Tab. 6 Impact of different hybrid guidance architecture updates of heuristic information on performance of CIFAR-10 dataset

对比方法	验证正确率/%		参数量/10⁶	搜索开销/GPU-Days
对比方法	搜索阶段	训练阶段	参数量/10⁶	搜索开销/GPU-Days
$I 1$	80.51	97.31	4.16	0.062
$I 2$	82.51	96.53	3.48	0.066
$I 3$	78.95	96.98	2.69	0.064
$I 1 + I 2$	84.71	97.31	2.70	0.068
$I 1 + I 3$	81.78	96.54	3.71	0.071
$I 2 + I 3$	84.42	97.07	2.53	0.070
$I 1 + I 2 + I 3$	87.61	97.55	3.77	0.120

Fig. 4 Comparison of verification accuracy when adding different loss items during search phase

Fig. 5 Heatmaps of architecture parameters for normal cells and reduce cells

References 55

[1]	ZHANG Y， DENG L， ZHU H， et al. Deep learning in food category recognition［J］. Information Fusion， 2023， 98： No.101859.
[2]	ASLANI S， JACOB J. Utilisation of deep learning for COVID-19 diagnosis［J］. Clinical Radiology， 2023， 78（2）： 150-157.
[3]	KOTWAL J， KASHYAP R， PATHAN S. Agricultural plant diseases identification： from traditional approach to deep learning［J］. Materials Today： Proceedings， 2023， 80（Pt 1）： 344-356.
[4]	YEH A H W， NORN C， KIPNIS Y， et al. De novo design of luciferases using deep learning［J］. Nature， 2023， 614（7949）： 774-780.
[5]	PUSHPARANI K， ROJA G， ANUSHA R， et al. Geological information extraction from satellite imagery using deep learning［C］// Proceedings of the 15th International Conference on Computing Communication and Networking Technologies. Piscataway： IEEE， 2024： 1-7.
[6]	POYSER M， BRECKON T P. Neural architecture search： a contemporary literature review for computer vision applications［J］. Pattern Recognition， 2024， 147： No.110052.
[7]	WHITE C， SAFARI M， SUKTHANKER R， et al. Neural architecture search： insights from 1000 papers［EB/OL］. ［2024-03-13］..
[8]	KANG J S， KANG J， KIM J J， et al. Neural architecture search survey： a computer vision perspective［J］. Sensors， 2023， 23（3）： No.1713.
[9]	WANG X， ZHU W. Advances in neural architecture search［J］. National Science Review， 2024， 11（8）： No.nwae282.
[10]	ZOPH B， LE Q V. Neural architecture search with reinforcement learning［EB/OL］. ［2025-01-23］..
[11]	PHAM H， GUAN M Y， ZOPH B， et al. Efficient neural architecture search via parameters sharing［C］// Proceedings of the 35th International Conference on Machine Learning. New York： JMLR.org， 2018： 4095-4104.
[12]	LIU H， SIMONYAN K， YANG Y. DARTS： differentiable architecture search［EB/OL］. ［2024-03-23］..
[13]	GUO Z， ZHANG X， MU H， et al. Single path one-shot neural architecture search with uniform sampling［C］// Proceedings of the 2020 European Conference on Computer Vision， LNCS 12361. Cham： Springer， 2020： 544-560.
[14]	CHU X， WANG X， ZHANG B， et al. DARTS-： robustly stepping out of performance collapse without indicators［EB/OL］. ［2024-05-03］..
[15]	LIANG H， ZHANG S， SUN J， et al. DARTS+： improved differentiable architecture search with early stopping［EB/OL］. ［2024-12-02］..
[16]	XU Y， XIE L， ZHANG X， et al. PC-DARTS： partial channel connections for memory-efficient architecture search［EB/OL］. ［2024-03-19］..
[17]	YE P， LI B， LI Y， et al. β-DARTS： Beta-decay regularization for differentiable architecture search［C］// Proceedings of the 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2022： 10864-10873.
[18]	REAL E， AGGARWAL A， HUANG Y， et al. Regularized evolution for image classifier architecture search［C］// Proceedings of the 33rd AAAI Conference on Artificial Intelligence. Palo Alto： AAAI Press， 2019： 4780-4789.
[19]	CHU X， ZHANG B， XU R. FairNAS： rethinking evaluation fairness of weight sharing neural architecture search［C］// Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2021： 12219-12228.
[20]	YOU S， HUANG T， YANG M， et al. GreedyNAS： towards fast one-shot NAS with greedy supernet［C］// Proceedings of the 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2020： 1996-2005.
[21]	BROCK A， LIM T， RITCHIE J M， et al. SMASH： one-shot model architecture search through hypernetworks［EB/OL］. ［2024-10-05］..
[22]	BENDER G， KINDERMANS P J， ZOPH B， et al. Understanding and simplifying one-shot architecture search［C］// Proceedings of the 35th International Conference on Machine Learning. New York： JMLR.org， 2018： 550-559.
[23]	LU Z， WHALEN I， BODDETI V， et al. NSGA-Net： neural architecture search using multi-objective genetic algorithm［C］// Proceedings of the 2019 Genetic and Evolutionary Computation Conference. New York： ACM， 2019： 419-427.
[24]	HUANG J， XUE B， SUN Y， et al. Particle swarm optimization for compact neural architecture search for image classification［J］. IEEE Transactions on Evolutionary Computation， 2023， 27（5）： 1298-1312.
[25]	HUANG J， XUE B， SUN Y， et al. Split-level evolutionary neural architecture search with elite weight inheritance［J］. IEEE Transactions on Neural Networks and Learning Systems， 2024， 35（10）： 13523-13537.
[26]	SUN Y， WANG H， XUE B， et al. Surrogate-assisted evolutionary deep learning using an end-to-end random forest-based performance predictor［J］. IEEE Transactions on Evolutionary Computation， 2020， 24（2）： 350-364.
[27]	WANG B， XUE B， ZHANG M. Surrogate-assisted particle swarm optimization for evolving variable-length transferable blocks for image classification［J］. IEEE Transactions on Neural Networks and Learning Systems， 2022， 33（8）： 3727-3740.
[28]	MARINI F， WALCZAK B. Particle Swarm Optimization （PSO）： a tutorial［J］. Chemometrics and Intelligent Laboratory Systems， 2015， 149（Pt B）： 153-165.
[29]	LIU Y， SUN Y， XUE B， et al. A survey on evolutionary neural architecture search［J］. IEEE Transactions on Neural Networks and Learning Systems， 2023， 34（2）： 550-570.
[30]	ZHANG M， SU S W， PAN S， et al. iDARTS： Differentiable architecture search with stochastic implicit gradients［C］// Proceedings of the 38th International Conference on Machine Learning. New York： JMLR.org， 2021： 12557-12566.
[31]	LOSHCHILOV I， HUTTER F. CMA-ES for hyperparameter optimization of deep neural networks［EB/OL］. ［2024-05-12］..
[32]	MILLETARI F， RIEKE N， BAUST M， et al. CFCM： segmentation via coarse to fine context memory［C］// Proceedings of the 2018 Medical Image Computing and Computer-Assisted Intervention， LNCS 11073. Cham： Springer， 2018： 667-674.
[33]	SARENI B， KRAHENBUHL L. Fitness sharing and niching methods revisited［J］. IEEE Transactions on Evolutionary Computation， 1998， 2（3）： 97-106.
[34]	ZHENG X， JI R， TANG L， et al. Multinomial distribution learning for effective neural architecture search［C］// Proceedings of the 2019 IEEE/CVF International Conference on Computer Vision. Piscataway： IEEE， 2019： 1304-1313.
[35]	ZHOU Y， XIE X， KUNG S Y. Exploiting operation importance for differentiable neural architecture search［J］. IEEE Transactions on Neural Networks and Learning Systems， 2022， 33（11）： 6235-6248.
[36]	DONG X， YANG Y. NAS-Bench-201： extending the scope of reproducible neural architecture search［EB/OL］. ［2024-05-11］..
[37]	ZOPH B， VASUDEVAN V， SHLENS J， et al. Learning transferable architectures for scalable image recognition［C］// Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2018： 8697-8710.
[38]	CAI H， ZHU L， HAN S. ProxylessNAS： direct neural architecture search on target task and hardware［EB/OL］. ［2025-02-14］..
[39]	CHEN X， XIE L， WU J， et al. Progressive DARTS： bridging the optimization gap for NAS in the wild［J］. International Journal of Computer Vision， 2021， 129（3）： 638-655.
[40]	XIE S， ZHENG H， LIU C， et al. SNAS： stochastic neural architecture search［EB/OL］. ［2024-12-04］..
[41]	XUE Y， QIN J. Partial connection based on channel attention for differentiable neural architecture search［J］. IEEE Transactions on Industrial Informatics， 2023， 19（5）： 6804-6813.
[42]	XUE Y， LU C， NERI F， et al. Improved differentiable architecture search with multi-stage progressive partial channel connections［J］. IEEE Transactions on Emerging Topics in Computational Intelligence， 2024， 8（1）： 32-43.
[43]	XUE Y， HAN X， WANG Z. Self-adaptive weight based on dual-attention for differentiable neural architecture search［J］. IEEE Transactions on Industrial Informatics， 2024， 20（4）： 6394-6403.
[44]	FANG X， XIE W， LI H， et al. DARTS-EAST： an edge-adaptive selection with topology first differentiable architecture selection method［J］. Applied Intelligence， 2025， 55（7）： No.526.
[45]	HAO D， PEI S. MIG-DARTS： towards effective differentiable architecture search by gradually mitigating the initial-channel gap between search and evaluation［J］. Neural Computing and Applications， 2025， 37（8）： 6085-6096.
[46]	WEI C， NIU C， TANG Y， et al. NPENAS： neural predictor guided evolution for neural architecture search［J］. IEEE Transactions on Neural Networks and Learning Systems， 2023， 34（11）： 8441-8455.
[47]	DONG J， HOU B， FENG L， et al. A cell-based fast memetic algorithm for automated convolutional neural architecture design［J］. IEEE Transactions on Neural Networks and Learning Systems， 2023， 34（11）： 9040-9053.
[48]	ZHANG H， JIN Y， HAO K. Evolutionary search for complete neural network architectures with partial weight sharing［J］. IEEE Transactions on Evolutionary Computation， 2022， 26（5）： 1072-1086.
[49]	ZHANG H， JIN Y， CHENG R， et al. Efficient evolutionary search of attention convolutional networks via sampled training and node inheritance［J］. IEEE Transactions on Evolutionary Computation， 2021， 25（2）： 371-385.
[50]	YUAN G， WANG B， XUE B， et al. Particle swarm optimization for efficiently evolving deep convolutional neural networks using an autoencoder-based encoding strategy［J］. IEEE Transactions on Evolutionary Computation， 2024， 28（5）： 1190-1204.
[51]	JIANG P， XUE Y， NERI F. Score predictor-assisted evolutionary neural architecture search［J］. IEEE Transactions on Emerging Topics in Computational Intelligence， 2025（Early Access）： 1-15.
[52]	YU H， PENG H， HUANG Y， et al. Cyclic differentiable architecture search［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2023， 45（1）： 211-228.
[53]	HU L， WANG Z， LI H， et al. ℓ-DARTS： light-weight differentiable architecture search with robustness enhancement strategy［J］. Knowledge-Based Systems， 2024， 288： No.111466.
[54]	XU Y， WANG Y， HAN K， et al. ReNAS： relativistic evaluation of neural architecture search［C］// Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway： IEEE， 2021： 4409-4418.
[55]	FAN L， WANG H. Surrogate-assisted evolutionary neural architecture search with network embedding［J］. Complex and Intelligent Systems， 2023， 9（3）： 3313-3331.

Neural network architecture search algorithm guided by hybrid heuristic information

混合启发信息指导神经网络架构搜索算法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 11

References 55

Related Articles 6

Recommended Articles

Metrics

[1]	Ye TIAN, Jinjin CHEN, Xingyi ZHANG. Hybrid optimizer combining evolutionary computation and gradient descent for constrained multi-objective optimization [J]. Journal of Computer Applications, 2024, 44(5): 1386-1392.
[2]	Zhihui GAO, Meng HAN, Shujuan LIU, Ang LI, Dongliang MU. Survey of high utility itemset mining methods based on intelligent optimization algorithm [J]. Journal of Computer Applications, 2023, 43(6): 1676-1686.
[3]	Wei LI, Yaochi FAN, Qiaoyong JIANG, Lei WANG, Qingzheng XU. Variable convolutional autoencoder method based on teaching-learning-based optimization for medical image classification [J]. Journal of Computer Applications, 2022, 42(2): 592-598.
[4]	WANG Liu-zheng HE Zhen-feng. Evolutionary K-means algorithm based on global splitting operator [J]. Journal of Computer Applications, 2012, 32(11): 3005-3008.
[5]	GUO Jin-cui ZOU Jin-xin SUN Peng-hui ZHUANG Yan. Improvement of dynamic multi-objective evolutionary and orthogonal test for four-branch satellite antenna [J]. Journal of Computer Applications, 2011, 31(10): 2880-2882.
[6]	ZHANG ZongFei. Quantum evolutionary algorithm for optimizing network intrusion signature database [J]. Journal of Computer Applications, 2010, 30(8): 2142-2145.