TL和DL结合论文整理(不断更新中)

本文汇总了多个关于深度学习及迁移学习的研究成果,探讨了这些技术如何应用于场景识别、视觉跟踪等多个领域,并介绍了通过深度网络进行特征提取和迁移的具体方法。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

1,Mesnil G, Rifai S, Bordes A, et al. Unsupervised Learning of Semantics of Object

 Detections for Scene Categorization[M]//Pattern Recognition Applications and

 Methods. Springer International Publishing, 2015: 209-224.


应用领域:场景识别


Keywords Unsupervised learning · Transfer learning · Deep learning · Scene categorization · Object detection


2,Wang N, Li S, Gupta A, et al. Transferring Rich Feature Hierarchies for Robust 

Visual Tracking[J]. arXiv preprint arXiv:1501.04587, 2015.


应用领域:visual tracking

CNN+TL


3,Srivastava N, Salakhutdinov R R. Multimodal learning with deep boltzmann

 machines[C]//Advances in neural information processing systems. 2012: 2222-2230.


实验室网站(有code)

http://www.cs.toronto.edu/~rsalakhu/publications.html 


4,Analysis using Progressively Trained and Domain Transferred Deep

 Networks[C]//The Twenty-Ninth AAAI Conference on Artificial Intelligence (AAAI).

 2015.You Q, Luo J, Jin H, et al. Robust Image Sentiment Ana


5,Oquab M, Bottou L, Laptev I, et al. Learning and transferring mid-level image

 representations using convolutional neural networks[C]//Computer Vision and Pattern

 Recognition (CVPR), 2014 IEEE Conference on. IEEE, 2014: 1717-1724.


6,Glorot X, Bordes A, Bengio Y. Domain adaptation for large-scale sentiment

 classification: A deep learning approach[C]//Proceedings of the 28th International

 Conference on Machine Learning (ICML-11). 2011: 513-520.


7,Chopra S, Balakrishnan S, Gopalan R. Dlid: Deep learning for domain adaptation

 by interpolating between domains[C]//ICML workshop on challenges in representation

 learning. 2013, 2: 5.


8,Bengio Y. Deep learning of representations for unsupervised and transfer

 learning[J]. Unsupervised and Transfer Learning Challenges in Machine Learning,

 Volume 7, 2012: 19.


补充。。。。


评论 2
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值