DDMI: a model information evaluation method based on deep dream✩

Yana Yang; Haoyu Li; Shuai Xiao; Guipeng Lan; Yong Zhu; Ziqiang Huo

doi:10.1016/j.dcan.2025.03.008

›› 2026, Vol. 12 ›› Issue (3) :530 -539. DOI: 10.1016/j.dcan.2025.03.008

Research Article

research-article

DDMI: a model information evaluation method based on deep dream^✩

Author information +

^a School of Electrical and Information Engineering, Tianjin University , Tianjin 300072, China

^b Tianjin Institute Of Software Engineering , Tianjin 300387, China

^* E-mail addresses: yangyana@tju.edu.cn (Y. Yang),

li_hy18@tju.edu.cn (H. Li),

xs611@tju.edu.cn (S. Xiao),

lgp@tju.edu.cn (G. Lan),

yongzhu@tjise.edu.cn (Y. Zhu),

huoziqiang@tju.edu.cn (Z. Huo).

Yana Yang received her bachelor’s degree in Computer Science and Technology from Tianjin University, China, in 2005. Her research interests include Software engineering, software information system project management, information fusion of marine hydrological and meteorological data, information visualization technology, and marine big data.

Haoyu Li received his bachelor’s degree from Tianjin University, China, in 2022. He received his Master’s degree from Tianjin University in 2025. His research interests include Deep learning, image processing, XAI.

Shuai Xiao received his bachelor’s degree from Zhengzhou University, China, in 2017. He received his Ph.D. from Tianjin University in 2022. His research interests include Image processing, computer vision, and image forensics.

Guipeng Lan graduated from Tianjin University in 2020 with a bachelor’s degree. He received his Ph.D. from Tianjin University in 2025. His research interests include Deep learning, AI for Science, Medicine image processing, Effective artificial intelligence, XAI.

Yong Zhu graduated from the Army Artillery Command Academy of the Chinese People’s Liberation Army in 2006 with a bachelor’s degree. He received his Ph.D. from Tianjin University. His research interests include Artificial Intelligence, Generative Adversarial Networks, Reinforcement Learning, and Cloud Computing.

Ziqiang Huo received the B.S. and M.S. degrees in electronic and information engineering from Tianjin University, Tianjin, China, in 2018 and 2021, respectively. He is currently working toward the Ph.D. degree in electronic and information systems with the School of Electrical and Information Engineering, Tianjin University, China. His research interests include Marine exploration, pattern recognition, image processing, artificial intelligence, big data computing.

Show less

History +

PDF (3073KB)

Abstract

In recent years, the Internet of Everything (IoE) has been developing rapidly; however, there are currently issues with efficiency and sustainability within IoE. To address this problem, model lightweighting can be employed by constraining the size of deep learning models, thereby reducing the demand for computational resources and ensuring the efficiency and sustainability of IoE devices. In this regard, we propose a model information evaluation method based on DeepDream. This method does not require real samples to participate; instead, it evaluates the importance of each neuron based on the model’s own information. Additionally, we introduce a method for automatically selecting high-information neurons, which can identify neurons that have a significant impact on the model. We also present a visualization method for neuron class information, which can visualize information related to various classes within neurons. Through experiments, we demonstrate the effectiveness of the methods we have proposed.

Keywords

Deep learning / Deep dream / Model information / IoE

Cite this article

Download citation ▾

Yana Yang, Haoyu Li, Shuai Xiao, Guipeng Lan, Yong Zhu, Ziqiang Huo. DDMI: a model information evaluation method based on deep dream^✩. , 2026, 12 (3) : 530-539 DOI:10.1016/j.dcan.2025.03.008

登录浏览全文

4963

注册一个新账户忘记密码

CRediT authorship contribution statement

Yana Yang: Writing -- review & editing, Software. Haoyu Li: Writing – original draft, Software, Methodology. Shuai Xiao: Writing -- review & editing, Funding acquisition. Guipeng Lan: Writing -- review & editing, Software, Methodology. Yong Zhu: Writing -- review & editing, Supervision, Software. Ziqiang Huo: Writing -- review & editing, Validation, Supervision.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China under Grant 62301356, and Joint Fund of Ministry of Education for Equipment Pre-research under Grant 8091B032254.

References

Publishing order | Descend order by publishing year | Descend order by cited within

[1]	International Telecommunication Union, ITU Internet Report 2005: The Internet of Things, in: Proceedings of the Workshop Report, International Telecommunication Union, 2005.

[2]	J. Yang, C. Cheng, S. Xiao, et al., High fidelity face—swapping with style convtransformer and latent space selection, IEEE Trans. Multimed. 26 (2024) 3604-3615.

[3]	D. Khurana, A. Koli, K. Khatter, et al., Natural language processing: state of the art, current trends and challenges, Multimed. Tools Appl. 82 (3) (2023) 3713-3744.

[4]	J. Yang, Z. Zhang, S. Xiao, et al., Efficient data—driven behavior identification based on vision transformers for human activity understanding, Neurocomputing 530 (2023) 104-115.

[5]	M. Xi, H. Dai, J. He, et al., A lightweight reinforcement—learning—based real—time path—planning method for unmanned aerial vehicles, IEEE Internet Things J. 11 (12) (2024) 21061-21071.

[6]	A. Rejeb, K. Rejeb, H. Treiblmaier, et al., The Internet of Things (IoT) in healthcare: taking stock and moving forward, Internet of Things 22 (2023) 100721.

[7]	J. Yang, S. Ma, Z. Zhang, et al., Say no to redundant information: unsupervised redundant feature elimination for active learning, IEEE Trans. Multimed. 26 (2024) 7721-7733.

[8]	OpenAI, GPT—4 technical report, arXiv:2303.08774, 2023.

[9]	A. Kirillov, E. Mintun, N. Ravi, et al., arXiv:2304.02643, 2023.

[10]	Y. Lecun, Optimal brain damage, Neural Inf. Proc. Syst. 2 (279) (1990) 598-605.

[11]	A. Islam, A.A. Amin, S.Y. Shin, FBI: a federated learning—based blockchain—embedded data accumulation scheme using drones for Internet of Things, IEEE Wirel. Commun. Lett. 11 (5) (2022) 972-976.

[12]	Y. Otoum, V. Chamola, A. Nayak, Federated and transfer learning—empowered intrusion detection for IoT applications, IEEE Internet Things Mag. 5 (3) (2022) 50-54.

[13]	A. Sharma, H. Babbar, Evaluation and analysis: Internet of Things using machine learning algorithms for detection of DDoS attacks, in: 2023 International Conference on Intelligent and Innovative Technologies in Computing, Electrical and Electronics (IITCEE), 2023, pp. 1203-1208.

[14]	W. Li, Q. Zhang, S. Deng, et al., Q—learning improved lightweight consensus algorithm for blockchain—structured Internet of Things, IEEE Internet Things J. 11 (2) (2024) 2855-2869.

[15]	A. Mohammadian, H. Zarrabi, S. Jabbehdari, et al., The effect of task processing management on energy consumption at the edge of Internet of things network with using reinforcement learning method, Comput. Ind. Eng. 195 (2024) 110426.

[16]	K. Simonyan, A. Vedaldi, A. Zisserman, Deep inside convolutional networks: visualising image classification models and saliency maps, arXiv:1312.6034, 2013.

[17]	B. Zhou, A. Khosla, A. Lapedriza, et al., Learning deep features for discriminative localization, in: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 2921-2929.

[18]	R.R. Selvaraju, M. Cogswell, A. Das, et al., Grad—CAM: visual explanations from deep networks via gradient—based localization, in: 2017 IEEE International Conference on Computer Vision (ICCV), 2017, pp. 618-626.

[19]	A. Chattopadhay, A. Sarkar, P. Howlader, et al., Grad—CAM++: generalized gradient—based visual explanations for deep convolutional networks, in: 2018 IEEE Winter Conference on Applications of Computer Vision (WACV), 2018, pp. 839-847.

[20]	B. Zhang, X. Gao, S. Yu, et al., Enhanced online CAM: single—stage weakly supervised semantic segmentation via collaborative guidance, Pattern Recognit. 156 (2024) 110787.

[21]	R. Xu, C. Wang, S. Xu, et al., Wave—like class activation map with representation fusion for weakly—supervised semantic segmentation, IEEE Trans. Multimed. 26 (2024) 581-592.

[22]	A. Mordvintsev, C. Olah, M. Tyka, Inceptionism: Going deeper into neural networks, Google research blog 20 (14) (2015) 5.

[23]	H. Yin, P. Molchanov, J.M. Alvarez, et al., Dreaming to distill: data—free knowledge transfer via deepinversion, in: 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020, pp. 8712-8721.

[24]	A. Singh, G. Gupta, R. Kapila, et al., CoDream: exchanging dreams instead of models for federated aggregation with heterogeneous models, arXiv:2402.15968, 2024.

[25]	C. Bucila, R. Caruana, A. Niculescu—Mizil, Model compression, in: Proceedings of the Twelfth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2006.

[26]	G. Hinton, O. Vinyals, J. Dean, Distilling the knowledge in a neural network, Comput. Sci. 14 (7) (2015) 38-39.

[27]	Z. Huang, N. Wang, Like what you like: knowledge distill via neuron selectivity transfer, arXiv:1707.01219, 2017.

[28]	R.G. Lopes, S. Fenu, T. Starner, Data—free knowledge distillation for deep neural networks, arXiv:1710.07535, 2017.

[29]	I. Goodfellow, J. Pouget—Abadie, M. Mirza, et al., Generative Adversarial Nets, MIT Press, 2014.

[30]	H. Chen, Y. Wang, C. Xu, et al., Data—free learning of student networks, in: 2019 IEEE/CVF International Conference on Computer Vision (ICCV), 2019, pp. 3513-3521.

[31]	J. Kim, M. Hyun, I. Chung, et al., Feature fusion for online mutual knowledge distillation, in: 2020 25th International Conference on Pattern Recognition (ICPR), 2021, pp. 4619-4625.

[32]	T. Liu, K.—M. Lam, R. Zhao, et al., Deep cross—modal representation learning and distillation for illumination—invariant pedestrian detection, IEEE Trans. Circuits Syst. Video Technol. 32 (1) (2022) 315-329.

[33]	T. Su, J. Zhang, Z. Yu, et al., STKD: distilling knowledge from synchronous teaching for efficient model compression, IEEE Trans. Neural Netw. Learn. Syst. 34 (12) (2023) 10051-10064.

[34]	W. Li, T. Qian, X. Li, et al., Adversarial multi—teacher distillation for semi—supervised relation extraction, IEEE Trans. Neural Netw. Learn. Syst. 35 (8) (2024) 11291-11301.

[35]	Y. LeCun, J. Denker, S. Solla, Optimal Brain Damage, Adv. Neural Inf. Process. Syst. 2 (1989).

[36]	H. Hu, R. Peng, Y.—W. Tai, et al., Network trimming: a data—driven neuron pruning approach towards efficient deep architectures, arXiv:1607.03250, 2016.

[37]	J.—H. Luo, J. Wu, An entropy—based pruning method for CNN compression, arXiv:1706.05791, 2017.

[38]	Y. He, G. Kang, X. Dong, et al., Soft filter pruning for accelerating deep convolutional neural networks, arXiv:1808.06866, 2018.

[39]	Y. Tu, M. Wang, S. Wang, A data—driven deep neural network pruning approach towards efficient digital signal modulation recognition, in: 2019 IEEE Globecom Workshops (GC Wkshps), 2019, pp. 1-5.

[40]	J. Tang, M. Liu, N. Jiang, et al., Data—free network pruning for model compression, in: 2021 IEEE International Symposium on Circuits and Systems (ISCAS), 2021, pp. 1-5.

[41]	W. Zou, Y. Wang, X. Fu, et al., Dreaming to prune image deraining networks, in: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 6013-6022.

[42]	X. Wang, Z. Zheng, Y. He, et al., Progressive local filter pruning for image retrieval acceleration, IEEE Trans. Multimed. 25 (2023) 9597-9607.

[43]	J. Deng, W. Dong, R. Socher, et al., ImageNet: a large—scale hierarchical image database, in: 2009 IEEE Conference on Computer Vision and Pattern Recognition, 2009, pp. 248-255.

[44]	K. He, X. Zhang, S. Ren, et al., Deep residual learning for image recognition, in: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016, pp. 770-778.

[45]	A. Krizhevsky, G. Hinton, et al., Learning multiple layers of features from tiny images, University of Toronto, Technical Report, 2009.

[46]	H. Li, A. Kadav, I. Durdanovic, et al., Pruning filters for efficient convnets, arXiv:1608.08710, 2017.

[47]	M. Lin, L. Cao, Y. Zhang, et al., Pruning networks with cross—layer ranking & k—reciprocal nearest filters, IEEE Trans. Neural Netw. Learn. Syst. 34 (11) (2023) 9139-9148.