{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2022,6,16]],"date-time":"2022-06-16T23:13:30Z","timestamp":1655421210108},"reference-count":33,"publisher":"MIT Press - Journals","issue":"7","content-domain":{"domain":["direct.mit.edu"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2022,6,16]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Sparse coding has been proposed as a theory of visual cortex and as an unsupervised algorithm for learning representations. We show empirically with the MNIST data set that sparse codes can be very sensitive to image distortions, a behavior that may hinder invariant object recognition. A locally linear analysis suggests that the sensitivity is due to the existence of linear combinations of active dictionary elements with high cancellation. A nearest-neighbor classifier is shown to perform worse on sparse codes than original images. For a linear classifier with a sufficiently large number of labeled examples, sparse codes are shown to yield higher accuracy than original images, but no higher than a representation computed by a random feedforward net. Sensitivity to distortions seems to be a basic property of sparse codes, and one should be aware of this property when applying sparse codes to invariant object recognition.<\/jats:p>","DOI":"10.1162\/neco_a_01513","type":"journal-article","created":{"date-parts":[[2022,6,7]],"date-time":"2022-06-07T20:37:36Z","timestamp":1654634256000},"page":"1616-1635","update-policy":"http:\/\/dx.doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":0,"title":["Sensitivity of Sparse Codes to Image Distortions"],"prefix":"10.1162","volume":"34","author":[{"given":"Kyle","family":"Luther","sequence":"first","affiliation":[{"name":"Department of Physics and Neuroscience Institute, Princeton University, Princeton, NJ 08544, U.S.A. kluther@princeton.edu"}]},{"given":"H. Sebastian","family":"Seung","sequence":"additional","affiliation":[{"name":"Neuroscience Institute and Department of Computer Science, Princeton University, Princeton, NJ 08544, U.S.A. sseung@princeton.edu"}]}],"member":"281","published-online":{"date-parts":[[2022,6,16]]},"reference":[{"key":"2022061622485015900_B1","doi-asserted-by":"publisher","first-page":"183","DOI":"10.1137\/080716542","article-title":"A fast iterative shrinkage-thresholding algorithm for linear inverse problems","volume":"2","author":"Beck","year":"2009","journal-title":"SIAM J. Imaging Sciences"},{"issue":"4","key":"2022061622485015900_B2","doi-asserted-by":"publisher","first-page":"827","DOI":"10.1162\/NECO_a_00247","article-title":"Learning intermediate-level representations of form and motion from natural movies","volume":"24","author":"Cadieu","year":"2012","journal-title":"Neural Computation"},{"issue":"9\u201310","key":"2022061622485015900_B3","doi-asserted-by":"publisher","first-page":"589","DOI":"10.1016\/j.crma.2008.03.014","article-title":"The restricted isometry property and its implications for compressed sensing","volume":"346","author":"Candes","year":"2008","journal-title":"Comptes rendus mathematique"},{"key":"2022061622485015900_B4","first-page":"9912","volume-title":"Advances in neural information processing systems","author":"Caron","year":"2020"},{"key":"2022061622485015900_B5","first-page":"1597","article-title":"A simple framework for contrastive learning of visual representations","volume-title":"Proceedings of the International Conference on Machine Learning","author":"Chen","year":"2020"},{"key":"2022061622485015900_B6","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-04167-0","volume-title":"Advances in neural information processing systems","author":"Chen","year":"2018"},{"issue":"6","key":"2022061622485015900_B7","doi-asserted-by":"publisher","first-page":"141","DOI":"10.1109\/MSP.2012.2211477","article-title":"The MNIST database of handwritten digit images for machine learning research","volume":"29","author":"Deng","year":"2012","journal-title":"IEEE Signal Processing Magazine"},{"key":"2022061622485015900_B8","doi-asserted-by":"publisher","first-page":"415","DOI":"10.1016\/j.neuron.2012.01.010","article-title":"How does the brain solve visual object recognition?","volume":"73","author":"DiCarlo","year":"2012","journal-title":"Neuron"},{"key":"2022061622485015900_B9","volume-title":"Advances in neural information processing systems","author":"Dosovitskiy","year":"2014"},{"issue":"12","key":"2022061622485015900_B10","doi-asserted-by":"publisher","first-page":"3736","DOI":"10.1109\/TIP.2006.881969","article-title":"Image denoising via sparse and redundant representations over learned dictionaries","volume":"15","author":"Elad","year":"2006","journal-title":"IEEE Transactions on Image Processing"},{"issue":"4","key":"2022061622485015900_B11","doi-asserted-by":"publisher","first-page":"193","DOI":"10.1007\/BF00344251","article-title":"Neocognitron: A self organizing neural network model for a mechanism of pattern recognition unaffected by shift in position","volume":"36","author":"Fukushima","year":"1980","journal-title":"Biological Cybernetics"},{"key":"2022061622485015900_B12","volume-title":"Advances in neural information processing systems","author":"Goodfellow","year":"2014"},{"key":"2022061622485015900_B13","author":"Grill","year":"2020","journal-title":"Bootstrap your own latent: A new approach to self-supervised learning"},{"key":"2022061622485015900_B14","doi-asserted-by":"crossref","DOI":"10.1109\/ICPR.2002.1048251","article-title":"Analyzing non-negative matrix factorization for image classification","volume-title":"Proceedings of the International Conference on Pattern Recognition.","author":"Guillamet","year":"2002"},{"key":"2022061622485015900_B15","doi-asserted-by":"crossref","first-page":"9726","DOI":"10.1109\/CVPR42600.2020.00975","article-title":"Momentum contrast for unsupervised visual representation learning","volume-title":"Proceedings of the 2020 IEEE\/CVF Conference on Computer Vision and Pattern Recognition","author":"He","year":"2020"},{"issue":"1","key":"2022061622485015900_B16","doi-asserted-by":"crossref","DOI":"10.1371\/journal.pone.0081813","article-title":"Sparsity-regularized HMAX for visual recognition","volume":"9","author":"Hu","year":"2014","journal-title":"PloS One"},{"key":"2022061622485015900_B17","doi-asserted-by":"crossref","first-page":"2146","DOI":"10.1109\/ICCV.2009.5459469","article-title":"What is the best multi-stage architecture for object recognition?","volume-title":"Proceedings of the 2009 IEEE 12th International Conference on Computer Vision","author":"Jarrett","year":"2009"},{"key":"2022061622485015900_B18","volume-title":"Advances in neural information processing systems","author":"Kingma","year":"2014"},{"key":"2022061622485015900_B19","article-title":"Auto-encoding variational Bayes","volume-title":"Proceedings of the 2nd International Conference on Learning Representations","author":"Kingma","year":"2014"},{"key":"2022061622485015900_B20","doi-asserted-by":"publisher","first-page":"607","DOI":"10.1038\/381607a0","article-title":"Emergence of simple-cell receptive field properties by learning a sparse code for natural images","volume":"381","author":"Olshausen","year":"1996","journal-title":"Nature"},{"issue":"23","key":"2022061622485015900_B21","doi-asserted-by":"publisher","first-page":"3311","DOI":"10.1016\/S0042-6989(97)00169-7","article-title":"Sparse coding with an overcomplete basis set: A strategy employed by V1?","volume":"37","author":"Olshausen","year":"1997","journal-title":"Vision Research"},{"key":"2022061622485015900_B22","first-page":"4797","volume-title":"Advances in neural information processing systems","author":"Oord","year":"2016"},{"key":"2022061622485015900_B23","doi-asserted-by":"publisher","DOI":"10.1167\/jov.20.12.10","article-title":"Selectivity and robustness of sparse coding networks","volume":"20","author":"Paiton","year":"2020","journal-title":"Journal of Vision"},{"key":"2022061622485015900_B24","article-title":"Unsupervised representation learning with deep convolutional generative adversarial networks","author":"Radford","year":"2016","journal-title":"CoRR"},{"key":"2022061622485015900_B25","doi-asserted-by":"crossref","DOI":"10.1145\/1273496.1273592","volume-title":"Self-taught learning: Transfer learning from unlabeled data.","author":"Raina","year":"2007"},{"key":"2022061622485015900_B26","doi-asserted-by":"crossref","first-page":"234","DOI":"10.1007\/978-3-319-24574-4_28","article-title":"U-net: Convolutional networks for biomedical image segmentation","volume-title":"Medical Image Computing and Computer-Assisted Intervention \u2013 MICCAI 2015","author":"Ronneberger","year":"2015"},{"key":"2022061622485015900_B27","first-page":"2110","article-title":"Adversarial robustness of supervised sparse coding","volume-title":"Advances in neural information processing systems","author":"Sulam","year":"2020"},{"key":"2022061622485015900_B28","doi-asserted-by":"publisher","first-page":"1456","DOI":"10.1214\/13-EJS815","article-title":"The Lasso problem and uniqueness","volume":"7","author":"Tibshirani","year":"2013","journal-title":"Electronic Journal of Statistics"},{"key":"2022061622485015900_B29","first-page":"586","article-title":"Face recognition using eigenfaces","volume-title":"Proceedings. of the 1991 IEEE Computer Society Conference on Computer Vision and Pattern Recognition","author":"Turk","year":"1991"},{"key":"2022061622485015900_B30","doi-asserted-by":"publisher","DOI":"10.5281\/zenodo.4569691","article-title":"elasticdeform: Elastic deformations for N-dimensional images","author":"van Tulder","year":"2021"},{"issue":"12","key":"2022061622485015900_B31","article-title":"Stacked denoising autoencoders: Learning useful representations in a deep network with a local denoising criterion","volume":"11","author":"Vincent","year":"2010","journal-title":"Journal of Machine Learning Research"},{"key":"2022061622485015900_B32","doi-asserted-by":"crossref","first-page":"1794","DOI":"10.1109\/CVPR.2009.5206757","article-title":"Linear spatial pyramid matching using sparse coding for image classification","volume-title":"Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition","author":"Yang","year":"2009"},{"key":"2022061622485015900_B33","article-title":"Barlow twins: Self-supervised learning via redundancy reduction","author":"Zbontar","year":"2021","journal-title":"CoRR"}],"container-title":["Neural Computation"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/direct.mit.edu\/neco\/article-pdf\/34\/7\/1616\/2030419\/neco_a_01513.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/direct.mit.edu\/neco\/article-pdf\/34\/7\/1616\/2030419\/neco_a_01513.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,6,16]],"date-time":"2022-06-16T22:49:16Z","timestamp":1655419756000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/neco\/article\/34\/7\/1616\/111334\/Sensitivity-of-Sparse-Codes-to-Image-Distortions"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,6,16]]},"references-count":33,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2022,6,16]]},"published-print":{"date-parts":[[2022,6,16]]}},"URL":"https:\/\/doi.org\/10.1162\/neco_a_01513","relation":{},"ISSN":["0899-7667","1530-888X"],"issn-type":[{"value":"0899-7667","type":"print"},{"value":"1530-888X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2022,7]]},"published":{"date-parts":[[2022,6,16]]}}}