[1]张静宇,续欣莹,谢刚,等.基于弹性权重巩固与知识蒸馏的垃圾持续分类[J].智能系统学报,2023,18(4):878-885.[doi:10.11992/tis.202211023]
ZHANG Jingyu,XU Xinying,XIE Gang,et al.Continuous classification of garbage based on the elastic weightconsolidation and knowledge distillation[J].CAAI Transactions on Intelligent Systems,2023,18(4):878-885.[doi:10.11992/tis.202211023]
点击复制
ZHANG Jingyu,XU Xinying,XIE Gang,et al.Continuous classification of garbage based on the elastic weightconsolidation and knowledge distillation[J].CAAI Transactions on Intelligent Systems,2023,18(4):878-885.[doi:10.11992/tis.202211023]
基于弹性权重巩固与知识蒸馏的垃圾持续分类
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
18
期数:
2023年第4期
页码:
878-885
栏目:
吴文俊人工智能科学技术奖论坛
出版日期:
2023-07-15
- Title:
- Continuous classification of garbage based on the elastic weightconsolidation and knowledge distillation
- Keywords:
- domestic garbage; image classification; continuous learning; deep learning; knowledge distillation; regularization; temperature coefficient; generalization capability
- 分类号:
- TP391
- 摘要:
- 针对目前的垃圾分类方法仅仅是对固定类别的常见生活垃圾分类，无法满足垃圾类别数量增长带来的动态持续分类要求的问题，本文提出了一种弹性权重巩固与知识蒸馏（elastic weight consolidation and knowledge distillation，EWC-KD）垃圾持续分类方法。该方法通过EWC正则化损失函数和蒸馏损失函数增强模型的记忆能力，EWC正则化损失函数限制重要参数的更新范围，带有温度系数的蒸馏损失函数通过保护类别标签中携带的类别信息增强模型的泛化能力。在5个垃圾分类任务上进行实验，结果表明该方法的性能优于对比方法，可以在所有任务上保持较高的分类准确率和较低的后向转移值，能够增强垃圾分类系统的持续分类能力。
- Abstract:
- The current garbage classification methods focus on the classification of common domestic garbage of fixed classes, which cannot meet the dynamic and continuous classification requirements brought by the growth of the number of garbage classes. To solve this problem, the paper proposes an elastic weight consolidation and knowledge distillation (EWC-KD) continuous garbage classification method. The method enhances the memory ability of the model through EWC regularization loss function and distillation loss function. EWC regularization loss function limits the update range of important parameters, and the distillation loss function with temperature coefficient enhances the generalization ability of the model by protecting the class information carried in the class label. Experiments on five garbage classification tasks show that the performance of this method is better than that of the comparison method. Our method can maintain high classification accuracy and low backward transfer value on all tasks, and can enhance the continuous classification ability of the garbage classification system.
- 参考文献/References:
-
[1] TONG Yeqing, LIU Jiafa, LIU Sizhe. China is implementing “Garbage Classification” action[J]. Environmental pollution, 2020, 259: 113707.
[2] 李金玉, 陈晓雷, 张爱华, 等. 基于深度学习的垃圾分类方法综述[J]. 计算机工程, 2022, 48(2): 1–9
LI Jinyu, CHEN Xiaolei, ZHANG Aihua, et al. Survey of garbage classification methods based on deep learning[J]. Computer engineering, 2022, 48(2): 1–9
[3] KANG Zhuang, YANG Jie, LI Guilan, et al. An automatic garbage classification system based on deep learning[J]. IEEE access, 2020, 8: 140019–140029.
[4] 张涛. 基于GAICNet的垃圾识别分类检测网络[J]. 智能计算机与应用, 2022, 12(4): 47–53
ZHANG Tao. Garbage identification classification and detection method based on GAICNet[J]. Intelligent computer and applications, 2022, 12(4): 47–53
[5] 许玉蕊, 刘银华, 高鑫. 基于特征融合卷积神经网络的垃圾分类[J]. 自动化与仪表, 2021, 36(9): 11–16
XU Yurui, LIU Yinhua, GAO Xin. Garbage classification based on feature fusion convolutional neural network[J]. Automation & instrumentation, 2021, 36(9): 11–16
[6] 杨旺功, 赵一飞. 注意力机制与双线性网络的垃圾图像分类研究[J]. 计算机仿真, 2021, 38(12): 222–226
YANG Wanggong, ZHAO Yifei. Research on garbage image classification based on attention mechanism and bilinear network[J]. Computer simulation, 2021, 38(12): 222–226
[7] ZENG Ming, LU Xiangzhe, XU Wenkang, et al. PublicGarbageNet: a deep learning framework for public garbage classification[C]// 39th Chinese Control Conference. Piscataway: IEEE, 2020: 7200-7205.
[8] MENG Sha, ZHANG Ning, REN Yunwen. X-DenseNet: deep learning for garbage classification based on visual images[J]. Journal of physics:conference series, 2020, 1575(1): 012139.
[9] SHI Cuiping, XIA Ruiyang, WANG Liguo. A novel multi-branch channel expansion network for garbage image classification[J]. IEEE access, 2020, 8: 154436–154452.
[10] BIRCANO?LU C, ATAY M, BE?ER F, et al. RecycleNet: intelligent waste sorting using deep neural networks[C]//2018 Innovations in Intelligent Systems and Applications. Piscataway: IEEE, 2018: 1?7.
[11] MAO Weilung, CHEN Weichun, WANG Chien, et al. Recycling waste classification using optimized convolutional neural network[J]. Resources, conservation and recycling, 2021, 164: 105132.
[12] YANG Jianfei, ZENG Zhaoyang, WANG Kai, et al. GarbageNet: a unified learning framework for robust garbage classification[J]. IEEE transactions on artificial intelligence, 2021, 2(4): 372–380.
[13] DE LANGE M, ALJUNDI R, MASANA M, et al. A continual learning survey: defying forgetting in classification tasks[J]. IEEE transactions on pattern analysis and machine intelligence, 2022, 44(7): 3366–3385.
[14] REBUFFI S A, KOLESNIKOV A, SPERL G, et al. iCaRL: incremental classifier and representation learning[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 5533?5542.
[15] SHIN H, LEE J K, KIM J, et al. Continual learning with deep generative replay[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York: ACM, 2017: 2994?3003.
[16] LOPEZ-PAZ D, RANZATO M. Gradient episodic memory for continual learning[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York: ACM, 2017: 6470?6479.
[17] CHAUDHRY A, RANZATO M, ROHRBACH M, et al. Efficient lifelong learning with A-GEM[EB/OL]. (2018?12?02)[2022?11?16]. https://arxiv.org/abs/1812.00420
[18] MALLYA A, LAZEBNIK S. PackNet: adding multiple tasks to a single network by iterative pruning[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2018: 7765?7773.
[19] FERNANDO C, BANARSE D, BLUNDELL C, et al. PathNet: evolution channels gradient descent in super neural networks[EB/OL]. (2017?01?30)[2022?11?16].https://arxiv.org/abs/1701.08734
[20] ALJUNDI R, CHAKRAVARTY P, TUYTELAARS T. Expert gate: lifelong learning with a network of experts[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 7120-7129.
[21] KIRKPATRICK J, PASCANU R, RABINOWITZ N, et al. Overcoming catastrophic forgetting in neural networks[J]. Proceedings of the national academy of sciences of the United States of America, 2017, 114(13): 3521–3526.
[22] LIU Xialei, MASANA M, HERRANZ L, et al. Rotate your networks: better weight consolidation and less catastrophic forgetting[C]//2018 24th International Conference on Pattern Recognition. Piscataway: IEEE, 2018: 2262-2268.
[23] LEE S W, KIM J H, JUN J, et al. Overcoming catastrophic forgetting by incremental moment matching[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York: ACM, 2017: 4655?4665.
[24] LI Zhizhong, HOIEM D. Learning without forgetting[J]. IEEE transactions on pattern analysis and machine intelligence, 2018, 40(12): 2935–2947.
[25] RANNEN A, ALJUNDI R, BLASCHKO M B, et al. Encoder based lifelong learning[C]//2017 IEEE International Conference on Computer Vision. Piscataway: IEEE, 2017: 1329?1337.
[26] CHAUDHRY A, DOKANIA P K, AJANTHAN T, et al. Riemannian walk for incremental learning: understanding forgetting and intransigence[M]//Computer Vision - ECCV 2018. Cham: Springer International Publishing, 2018: 556?572.
[27] HINTON G, VINYALS O, DEAN J. Distilling the knowledge in a neural network[EB/OL]. (2015?03?09)[2022?11?16]. https://arxiv.org/abs/1503.02531.
[28] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2017, 60(6): 84–90.
[29] HUANG Gao, LIU Zhuang, VAN DER MAATEN L, et al. Densely connected convolutional networks[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2017: 2261?2269.
[30] HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 770?778.
[31] YAN Xin. Garbagenet: a large-scale garbage dataset for image classification[EB/OL]. (2023?01?20)[2023?03?10]. https://git.openi.org.cn/Garbage sorting/GarbageNet.
- 相似文献/References:
-
[1]李海峰,杜军平.颜色特征的图像分类技术研究[J].智能系统学报,2008,3(2):65.[doi:CNKI:SUN:ZNXT.0.2008-02-017]
[2]李海峰,杜军平.颜色特征的图像分类技术研究[J].智能系统学报,2008,3(2):155.
LI Hai-feng,DU Jun-ping.Image classification technology based on color features[J].CAAI Transactions on Intelligent Systems,2008,3():155.
[3]姚伏天,钱沄涛.高斯过程及其在高光谱图像分类中的应用[J].智能系统学报,2011,6(5):396.
YAO Futian,QIAN Yuntao.Gaussian process and its applications in hyperspectral image classification[J].CAAI Transactions on Intelligent Systems,2011,6():396.
[4]尤雅萍,成运,苏松志,等.基于谱域-空域结合特征和图割原理的高光谱图像分类[J].智能系统学报,2015,10(2):201.[doi:10.3969/j.issn.1673-4785.201410040]
YOU Yaping,CHENG Yun,SU Songzhi,et al.Hyperspectral image classification based on spectral-spatial combination features and graph cut[J].CAAI Transactions on Intelligent Systems,2015,10():201.[doi:10.3969/j.issn.1673-4785.201410040]
[5]赵骞,李敏,赵晓杰,等.基于感受野学习的特征词袋模型简化算法[J].智能系统学报,2016,11(5):663.[doi:10.11992/tis.201601001]
ZHAO Qian,LI Min,ZHAO Xiaojie,et al.Learning receptive fields for compact bag-of-feature model[J].CAAI Transactions on Intelligent Systems,2016,11():663.[doi:10.11992/tis.201601001]
[6]费宇杰,吴小俊.一种局部聚合描述符和组显著编码相结合的编码方法[J].智能系统学报,2017,12(2):172.[doi:10.11992/tis.201602010]
FEI Yujie,WU Xiaojun.A new feature coding algorithm based on the combination of group salient coding and VLAD[J].CAAI Transactions on Intelligent Systems,2017,12():172.[doi:10.11992/tis.201602010]
[7]杨梦铎,栾咏红,刘文军,等.基于自编码器的特征迁移算法[J].智能系统学报,2017,12(6):894.[doi:10.11992/tis.201706037]
YANG Mengduo,LUAN Yonghong,LIU Wenjun,et al.Feature transfer algorithm based on an auto-encoder[J].CAAI Transactions on Intelligent Systems,2017,12():894.[doi:10.11992/tis.201706037]
[8]马忠丽,刘权勇,武凌羽,等.一种基于联合表示的图像分类方法[J].智能系统学报,2018,13(2):220.[doi:10.11992/tis.201611036]
MA Zhongli,LIU Quanyong,WU Lingyu,et al.Syncretic representation method for image classification[J].CAAI Transactions on Intelligent Systems,2018,13():220.[doi:10.11992/tis.201611036]
[9]魏彩锋,孙永聪,曾宪华.图正则化字典对学习的轻度认知功能障碍预测[J].智能系统学报,2019,14(2):369.[doi:10.11992/tis.201709033]
WEI Caifeng,SUN Yongcong,ZENG Xianhua.Dictionary pair learning with graph regularization for mild cognitive impairment prediction[J].CAAI Transactions on Intelligent Systems,2019,14():369.[doi:10.11992/tis.201709033]
[10]赵玉新,赵廷.海底声呐图像智能底质分类技术研究综述[J].智能系统学报,2020,15(3):587.[doi:10.11992/tis.202004026]
ZHAO Yuxin,ZHAO Ting.Survey of the intelligent seabed sediment classification technology based on sonar images[J].CAAI Transactions on Intelligent Systems,2020,15():587.[doi:10.11992/tis.202004026]
备注/Memo
收稿日期:2022-11-16。
基金项目:山西省回国留学人员科研项目（2021-046）；山西省自然科学基金项目（202103021224056）；山西省科技合作交流专项（202104041101030）.
作者简介:张静宇，硕士研究生，主要研究方向为深度学习和图像处理;续欣莹，教授，美国圣荷西州立大学访问学者，中国人工智能学会科普工作委员会常务委员、认知系统与信息处理专委会委员，主要研究方向为人工智能、视觉感知与智能控制。主持国家级、省部级和企业等重要项目20余项，发表学术论文80余篇;刘华平，教授，博士生导师，国家级人才，中国人工智能学会理事、中国人工智能学会认知系统与信息处理专业委员会副主任，国家杰出青年科学基金获得者，主要研究方向为机器人感知、学习与控制，获吴文俊科技进步一等奖。主持国家自然科学基金重点 项目 2 项，发表学术论文 300 余篇
通讯作者:刘华平.E-mail:hpliu@tsinghua.edu.cn
更新日期/Last Update:
1900-01-01