Deep Learning Vector Quantization

Deep Learning Vector Quantization
复制标题

深度学习矢量量化

DOI:
--
复制
发表时间:
2016
期刊:
The European Symposium on Artificial Neural Networks
影响因子:
--
通讯作者:
Aaron C. Courville
Aaron C. Courville
中科院分区:
--
文献类型:
--
作者:
H. D. Vries;R. Memisevic;Aaron C. Courville

文献摘要

参考文献

被引文献

相似文献

.虽然深度神经网络(DNN)在图像识别任务上取得了令人印象深刻的性能,但之前的研究报告称,DNN对无法识别的图像给出了高置信度的预测。由于观察到这种欺骗性的例子可能是由log-softmax的外推性质引起的,我们建议将联合收割机神经网络与学习矢量量化(LVQ)结合起来。我们提出的方法称为Deep LVQ(DLVQ),在MNIST上实现了相当的性能,同时对欺骗和对抗性示例更具鲁棒性。
. While deep neural nets (DNN’s) achieve impressive performance on image recognition tasks, previous studies have reported that DNN’s give high confidence predictions for unrecognizable images. Motivated by the observation that such fooling examples might be caused by the extrapolating nature of the log-softmax, we propose to combine neural networks with Learning Vector Quantization (LVQ). Our proposed method, called Deep LVQ (DLVQ), achieves comparable performance on MNIST while being more robust against fooling and adversarial examples.
DOI: 10.1016/j.neunet.2011.10.001
发表时间: 2012-02-01
期刊: NEURAL NETWORKS
影响因子: 7.8
作者:
Bunte, Kerstin;Schneider, Petra;Biehl, Michael
通讯作者: Biehl, Michael