2020

How Much Position Information Do Convolutional Neural Networks Encode?

Islam, Md Amirul, Jia, Sen, Bruce, Neil D. B.

Understand

In contrast to fully connected networks, Convolutional Neural Networks (CNNs) achieve efficiency by learning weights associated with local filters with a finite spatial extent.

An implication of this is that a filter may know what it is looking at, but not where it is positioned in the image.
Information concerning absolute position is inherently useful, and it is reasonable to assume that deep CNNs may implicitly learn to encode this information if there is a means to do so.
In this paper, we test this hypothesis revealing the surprising degree of absolute position information that is encoded in commonly used neural networks.

Built on

Amulet: Aggregating multi-level convolutional features for salient object detection
Original
Pingping Zhang, Dong Wang, Huchuan Lu, Hongyu Wang, and Xiang Ruan · 2001
Earlier work this paper cites.
Unsupervised discovery of nonlinear structure using contrastive backpropagation
Geoffrey Hinton, Simon Osindero, Max Welling, and Yee-Whye Teh · 2006
Earlier work this paper cites.
Visualizing higher-layer features of a deep network
Dumitru Erhan, Yoshua Bengio, Aaron Courville, and Pascal Vincent · 2009
Earlier work this paper cites.
Understanding the difficulty of training deep feedforward neural networks
Xavier Glorot and Yoshua Bengio · 2010
Earlier work this paper cites.
Adaptive deconvolutional networks for mid and high level feature learning
M. D. Zeiler, G. W. Taylor, and R. Fergus · 2011
Earlier work this paper cites.
The secrets of salient object segmentation
Y. Li, X. Hou, C. Koch, J. M. Rehg, and A. L. Yuille · 2014
Earlier work this paper cites.
Very deep convolutional networks for large-scale image recognition
Original
Karen Simonyan and Andrew Zisserman · 2014
Earlier work this paper cites.
Deepface: Closing the gap to human-level performance in face verification
Yaniv Taigman, Ming Yang, Marc’Aurelio Ranzato, and Lior Wolf · 2014
Earlier work this paper cites.
Visualizing and understanding convolutional networks
Matthew D Zeiler and Rob Fergus · 2014
Earlier work this paper cites.
Global contrast based salient region detection
M. Cheng, N. J. Mitra, X. Huang, P. H. S. Torr, and S. Hu · 2015
Earlier work this paper cites.

Similar

Fully convolutional networks for semantic segmentation
Jonathan Long, Evan Shelhamer, and Trevor Darrell · 2015
Cited alongside, same era.
Learning deconvolution network for semantic segmentation
Hyeonwoo Noh, Seunghoon Hong, and Bohyung Han · 2015
Cited alongside, same era.
Faster r-cnn: Towards real-time object detection with region proposal networks
Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun · 2015
Cited alongside, same era.
A recurrent neural network based alternative to convolutional networks
Original
F Visin, K Kastner, K Cho, M Matteucci, A Courville, and Y Bengio · 2015
Cited alongside, same era.
Deep residual learning for image recognition
Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun · 2016
Cited alongside, same era.
You only look once: Unified, real-time object detection
Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi · 2016
Cited alongside, same era.
Understanding deep learning requires rethinking generalization
Original
Chiyuan Zhang, Samy Bengio, Moritz Hardt, Benjamin Recht, and Oriol Vinyals · 2016
Cited alongside, same era.
Segnet: A deep convolutional encoder-decoder architecture for scene segmentation
Vijay Badrinarayanan, Alex Kendall, and Roberto Cipolla · 2017
Cited alongside, same era.
Gated feedback refinement network for dense image labeling
Md Amirul Islam, Mrigank Rochan, Neil DB Bruce, and Yang Wang · 2017
Cited alongside, same era.
Dynamic routing between capsules
Sara Sabour, Nicholas Frosst, and Geoffrey E Hinton · 2017
Cited alongside, same era.

Then

Grad-cam: Visual explanations from deep networks via gradient-based localization
Ramprasaath R Selvaraju, Michael Cogswell, Abhishek Das, Ramakrishna Vedantam, Devi Parikh, and Dhruv Batra · 2017
Later among the works it cites.
Learning to detect salient objects with image-level supervision
Lijun Wang, Huchuan Lu, Yifan Wang, Mengyang Feng, Dong Wang, Baocai Yin, and Xiang Ruan · 2017
Later among the works it cites.
Deep convolutional networks do not classify based on global object shape
Nicholas Baker, Hongjing Lu, Gennady Erlikhman, and Philip J. Kellman · 2018
Later among the works it cites.
Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs
Liang-Chieh Chen, George Papandreou, Iasonas Kokkinos, Kevin Murphy, and Alan L Yuille · 2018
Later among the works it cites.
Predicting human eye fixations via an lstm-based saliency attentive model
M. Cornia, L. Baraldi, G. Serra, and R. Cucchiara · 2018
Later among the works it cites.
Revisiting salient object detection: Simultaneous detection, ranking, and subitizing of multiple salient objects
Md Amirul Islam, Mahmoud Kalash, and Neil DB Bruce · 2018
Later among the works it cites.
Eml-net: An expandable multi-layer network for saliency prediction
Original
Sen Jia and Neil DB Bruce · 2018
Later among the works it cites.
The mythos of model interpretability
Zachary C Lipton · 2018
Later among the works it cites.
Picanet: Learning pixel-wise contextual attention for saliency detection
Nian Liu, Junwei Han, and Ming-Hsuan Yang · 2018
Later among the works it cites.
Richer and deeper supervision network for salient object detection
Original
Sen Jia and Neil DB Bruce · 2019
Later among the works it cites.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

Reading the bibliography…