Added multiple references
Signed-off-by: Jim Martens <github@2martens.de>
This commit is contained in:
parent
e4d0883b18
commit
98d57633ff
118
ma.bib
118
ma.bib
|
|
@ -115,8 +115,8 @@ dropout as bayesian approximation},
|
|||
publisher = {Curran Associates, Inc.},
|
||||
pages = {1097--1105},
|
||||
url = {http://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdf},
|
||||
abstract = {We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes. On the test data, we achieved top-1 and top-5 error rates of 37.5%
|
||||
and 17.0% which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implementation of the convolution operation. To reduce overfitting in the fully-connected layers we employed a recently-developed regularization method called “dropout” that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.},
|
||||
abstract = {We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes. On the test data, we achieved top-1 and top-5 error rates of 37.5\%
|
||||
and 17.0\% which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implementation of the convolution operation. To reduce overfitting in the fully-connected layers we employed a recently-developed regularization method called “dropout” that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3\%, compared to 26.2\% achieved by the second-best entry.},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/07_imagenet-classification-with-deep-convolutional-neural-networks_krizhevsky.pdf:PDF},
|
||||
owner = {jim},
|
||||
timestamp = {2019.01.02},
|
||||
|
|
@ -690,7 +690,7 @@ to construct explicit models for non-normal classes. Application includes infere
|
|||
number = {11},
|
||||
pages = {847--856},
|
||||
doi = {10.3844/ajassp.2015.847.856},
|
||||
abstract = {Recent popularity of RGB-D sensors mostly comes from the fact that RGB-images and depth maps supplement each other in machine vision tasks, such as object detection and recognition. This article addresses a problem of RGB and depth data fusion for pedestrian detection. We propose pedestrian detection algorithm that involves fusion of outputs of 2D- and 3D-detectors based on deep autoencoders. Outputs are fused with neural network classifier trained using a dataset which entries are represented by pairs of reconstruction errors of 2D- and 3D-autoencoders. Experimental results show that fusing outputs almost totally eliminate false accepts (precision is 99.8%) and brings recall to 93.2% when tested on the combined dataset that includes a lot of samples with significantly distorted human silhouette. Though we use walking pedestrians as objects of interest, there are few pedestrian-specific processing blocks in this algorithm, so, in general, it can be applied to any type of objects.},
|
||||
abstract = {Recent popularity of RGB-D sensors mostly comes from the fact that RGB-images and depth maps supplement each other in machine vision tasks, such as object detection and recognition. This article addresses a problem of RGB and depth data fusion for pedestrian detection. We propose pedestrian detection algorithm that involves fusion of outputs of 2D- and 3D-detectors based on deep autoencoders. Outputs are fused with neural network classifier trained using a dataset which entries are represented by pairs of reconstruction errors of 2D- and 3D-autoencoders. Experimental results show that fusing outputs almost totally eliminate false accepts (precision is 99.8\%) and brings recall to 93.2\% when tested on the combined dataset that includes a lot of samples with significantly distorted human silhouette. Though we use walking pedestrians as objects of interest, there are few pedestrian-specific processing blocks in this algorithm, so, in general, it can be applied to any type of objects.},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/45_pedestrian-detection-in-rgbd-using-autoencoders.pdf:PDF},
|
||||
owner = {jim},
|
||||
publisher = {Science Publications},
|
||||
|
|
@ -727,4 +727,116 @@ to construct explicit models for non-normal classes. Application includes infere
|
|||
timestamp = {2019.08.13},
|
||||
}
|
||||
|
||||
@Article{MacKay1992,
|
||||
author = {David J. C. MacKay},
|
||||
title = {A Practical Bayesian Framework for Backpropagation Networks},
|
||||
journal = {Neural Computation},
|
||||
year = {1992},
|
||||
volume = {4},
|
||||
number = {3},
|
||||
pages = {448--472},
|
||||
doi = {10.1162/neco.1992.4.3.448},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/50_practical-bayesian-framework-for-backprop-networks.pdf:PDF},
|
||||
owner = {jim},
|
||||
publisher = {{MIT} Press - Journals},
|
||||
timestamp = {2019.08.26},
|
||||
}
|
||||
|
||||
@Book{Neal1996,
|
||||
author = {Radford M. Neal},
|
||||
title = {Bayesian Learning for Neural Networks},
|
||||
year = {1996},
|
||||
publisher = {Springer New York},
|
||||
doi = {10.1007/978-1-4612-0745-0},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/52_bayesian-learning-for-neural-networks.pdf:PDF},
|
||||
owner = {jim},
|
||||
timestamp = {2019.08.26},
|
||||
}
|
||||
|
||||
@Article{Teye2018,
|
||||
author = {Mattias Teye and Hossein Azizpour and Kevin Smith},
|
||||
title = {Bayesian Uncertainty Estimation for Batch Normalized Deep Networks},
|
||||
journal = {arXiv preprint},
|
||||
date = {2018-02-18},
|
||||
eprint = {1802.06455v2},
|
||||
eprintclass = {stat.ML},
|
||||
eprinttype = {arXiv},
|
||||
abstract = {We show that training a deep network using batch normalization is equivalent to approximate inference in Bayesian models. We further demonstrate that this finding allows us to make meaningful estimates of the model uncertainty using conventional architectures, without modifications to the network or the training procedure. Our approach is thoroughly validated by measuring the quality of uncertainty in a series of empirical experiments on different tasks. It outperforms baselines with strong statistical significance, and displays competitive performance with recent Bayesian approaches.},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/51_bayesian-uncertainty-estimation-batch-normalized-deep-networks.pdf:PDF},
|
||||
keywords = {stat.ML},
|
||||
owner = {jim},
|
||||
timestamp = {2019.08.26},
|
||||
}
|
||||
|
||||
@Article{Ghahramani2015,
|
||||
author = {Zoubin Ghahramani},
|
||||
title = {Probabilistic machine learning and artificial intelligence},
|
||||
journal = {Nature},
|
||||
year = {2015},
|
||||
volume = {521},
|
||||
number = {7553},
|
||||
pages = {452--459},
|
||||
doi = {10.1038/nature14541},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/53_Probabilistic-machine-learning-and-ai.pdf:PDF},
|
||||
owner = {jim},
|
||||
publisher = {Springer Science and Business Media {LLC}},
|
||||
timestamp = {2019.08.26},
|
||||
}
|
||||
|
||||
@Article{Ioffe2015,
|
||||
author = {Sergey Ioffe and Christian Szegedy},
|
||||
title = {Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift},
|
||||
journal = {arXiv preprint},
|
||||
date = {2015-03-02},
|
||||
eprint = {1502.03167v3},
|
||||
eprintclass = {cs.LG},
|
||||
eprinttype = {arXiv},
|
||||
abstract = {Training Deep Neural Networks is complicated by the fact that the distribution of each layer's inputs changes during training, as the parameters of the previous layers change. This slows down the training by requiring lower learning rates and careful parameter initialization, and makes it notoriously hard to train models with saturating nonlinearities. We refer to this phenomenon as internal covariate shift, and address the problem by normalizing layer inputs. Our method draws its strength from making normalization a part of the model architecture and performing the normalization for each training mini-batch. Batch Normalization allows us to use much higher learning rates and be less careful about initialization. It also acts as a regularizer, in some cases eliminating the need for Dropout. Applied to a state-of-the-art image classification model, Batch Normalization achieves the same accuracy with 14 times fewer training steps, and beats the original model by a significant margin. Using an ensemble of batch-normalized networks, we improve upon the best published result on ImageNet classification: reaching 4.9\% top-5 validation error (and 4.8\% test error), exceeding the accuracy of human raters.},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/54_batch_normalization.pdf:PDF},
|
||||
keywords = {cs.LG},
|
||||
owner = {jim},
|
||||
timestamp = {2019.08.26},
|
||||
}
|
||||
|
||||
@InProceedings{Li2019,
|
||||
author = {Li, Xiang and Chen, Shuo and Hu, Xiaolin and Yang, Jian},
|
||||
title = {Understanding the Disharmony Between Dropout and Batch Normalization by Variance Shift},
|
||||
booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
|
||||
year = {2019},
|
||||
pages = {2682--2690},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/55_Li_Understanding_the_Disharmony_Between_Dropout_and_Batch_Normalization_by_Variance_CVPR_2019_paper.pdf:PDF},
|
||||
owner = {jim},
|
||||
timestamp = {2019.08.26},
|
||||
}
|
||||
|
||||
@Article{Postels2019,
|
||||
author = {Janis Postels and Francesco Ferroni and Huseyin Coskun and Nassir Navab and Federico Tombari},
|
||||
title = {Sampling-free Epistemic Uncertainty Estimation Using Approximated Variance Propagation},
|
||||
journal = {arXiv preprint},
|
||||
date = {2019-08-21},
|
||||
eprint = {1908.00598v2},
|
||||
eprintclass = {cs.LG},
|
||||
eprinttype = {arXiv},
|
||||
abstract = {We present a sampling-free approach for computing the epistemic uncertainty of a neural network. Epistemic uncertainty is an important quantity for the deployment of deep neural networks in safety-critical applications, since it represents how much one can trust predictions on new data. Recently promising works were proposed using noise injection combined with Monte-Carlo sampling at inference time to estimate this quantity (e.g. Monte-Carlo dropout). Our main contribution is an approximation of the epistemic uncertainty estimated by these methods that does not require sampling, thus notably reducing the computational overhead. We apply our approach to large-scale visual tasks (i.e., semantic segmentation and depth regression) to demonstrate the advantages of our method compared to sampling-based approaches in terms of quality of the uncertainty estimates as well as of computational overhead.},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/56_sampling-free-epistemic-uncertainty-estimation-using-approximated-variance-propagation.pdf:PDF},
|
||||
keywords = {cs.LG, stat.ML},
|
||||
owner = {jim},
|
||||
timestamp = {2019.08.26},
|
||||
}
|
||||
|
||||
@Article{Luo2018,
|
||||
author = {Ping Luo and Xinjiang Wang and Wenqi Shao and Zhanglin Peng},
|
||||
title = {Towards Understanding Regularization in Batch Normalization},
|
||||
journal = {arXiv preprint},
|
||||
date = {2018-09-04},
|
||||
eprint = {http://arxiv.org/abs/1809.00846v4},
|
||||
eprintclass = {cs.LG},
|
||||
eprinttype = {arXiv},
|
||||
abstract = {Batch Normalization (BN) improves both convergence and generalization in training neural networks. This work understands these phenomena theoretically. We analyze BN by using a basic block of neural networks, consisting of a kernel layer, a BN layer, and a nonlinear activation function. This basic network helps us understand the impacts of BN in three aspects. First, by viewing BN as an implicit regularizer, BN can be decomposed into population normalization (PN) and gamma decay as an explicit regularization. Second, learning dynamics of BN and the regularization show that training converged with large maximum and effective learning rate. Third, generalization of BN is explored by using statistical mechanics. Experiments demonstrate that BN in convolutional neural networks share the same traits of regularization as the above analyses.},
|
||||
file = {:/home/jim/Documents/Studium/MA/Literatur/57_towards-understanding-regularization-in-bn.pdf:PDF},
|
||||
keywords = {cs.LG, cs.CV, cs.SY, stat.ML},
|
||||
owner = {jim},
|
||||
timestamp = {2019.08.26},
|
||||
}
|
||||
|
||||
@Comment{jabref-meta: databaseType:biblatex;}
|
||||
|
|
|
|||
Loading…
Reference in New Issue