Split (1)

train · 309 rows

paper_id stringlengths 13 13	title stringlengths 14 123	abstract stringlengths 72 1.87k	full_text stringlengths 0 89.3k	authors stringlengths 10 110	decision stringclasses 1 value	year int64 2.01k 2.01k	api_raw_submission stringlengths 798 2.87k	review stringlengths 12 16.6k	reviewer_id stringlengths 2 110	rating null	confidence null	api_raw_review stringlengths 455 17.4k	criteria_count dict	reward_value float64 0 3.25	reward_value_length_adjusted float64 -303.12 2.54	length_penalty float64 0 305	reward_u float64 0 3.27	reward_h float64 0.01 1.63	meteor_score float64 0 0.18	criticism float64 0 1	example float64 0 0.5	importance_and_relevance float64 0 1	materials_and_methods float64 0 1	praise float64 0 0.67	presentation_and_reporting float64 0 1	results_and_discussion float64 0 1	suggestion_and_solution float64 0 1	dimension_scores dict	overall_score float64 0 3.25	source stringclasses 1 value	review_src stringclasses 1 value	conversations null
i87JIQTAnB8AQ	The Diagonalized Newton Algorithm for Nonnegative Matrix Factorization	Non-negative matrix factorization (NMF) has become a popular machine learning approach to many problems in text mining, speech and image processing, bio-informatics and seismic data analysis to name a few. In NMF, a matrix of non-negative data is approximated by the low-rank product of two matrices with non-negative en...	The Diagonalized Newton Algorithm for Non- negative Matrix Factorization Hugo Van hamme 1 University of Leuven, dept. ESAT 2 Kasteelpark Arenberg 10 – bus 2441, 3001 Leuven, B elgium 3 hugo.vanhamme@esat.kuleuven.be 4 Abstract 5 Non-negative matrix factorization (NMF) has become a popular machine 6 learn...	Hugo Van hamme	Unknown	2,013	{"id": "i87JIQTAnB8AQ", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 60, "content": {"title": "The Diagonalized Newton Algorithm for Nonnegative Matrix Factorization", "decision": "conferencePoster-iclr2013-confe...	[Review]: Summary: The paper presents a new algorithm for solving L1 regularized NMF problems in which the fitting term is the Kullback-Leiber divergence. The strategy combines the classic multiplicative updates with a diagonal approximation of Newton's method for solving the KKT conditions of the NMF optimization p...	anonymous reviewer 4322	null	null	{"id": "FFkZF49pZx-pS", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362210360000, "tmdate": 1362210360000, "ddate": null, "number": 4, "content": {"title": "review of The Diagonalized Newton Algorithm for Nonnegative Matrix Factorization", "review": "Summary:\r\n\r\nThe pape...	{ "criticism": 4, "example": 0, "importance_and_relevance": 4, "materials_and_methods": 12, "praise": 2, "presentation_and_reporting": 4, "results_and_discussion": 4, "suggestion_and_solution": 5, "total": 21 }	1.666667	0.656591	1.010076	1.707926	0.429151	0.04126	0.190476	0	0.190476	0.571429	0.095238	0.190476	0.190476	0.238095	{ "criticism": 0.19047619047619047, "example": 0, "importance_and_relevance": 0.19047619047619047, "materials_and_methods": 0.5714285714285714, "praise": 0.09523809523809523, "presentation_and_reporting": 0.19047619047619047, "results_and_discussion": 0.19047619047619047, "suggestion_and_solution": 0.23...	1.666667	iclr2013	openreview	null
i87JIQTAnB8AQ	The Diagonalized Newton Algorithm for Nonnegative Matrix Factorization	Non-negative matrix factorization (NMF) has become a popular machine learning approach to many problems in text mining, speech and image processing, bio-informatics and seismic data analysis to name a few. In NMF, a matrix of non-negative data is approximated by the low-rank product of two matrices with non-negative en...	The Diagonalized Newton Algorithm for Non- negative Matrix Factorization Hugo Van hamme 1 University of Leuven, dept. ESAT 2 Kasteelpark Arenberg 10 – bus 2441, 3001 Leuven, B elgium 3 hugo.vanhamme@esat.kuleuven.be 4 Abstract 5 Non-negative matrix factorization (NMF) has become a popular machine 6 learn...	Hugo Van hamme	Unknown	2,013	{"id": "i87JIQTAnB8AQ", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 60, "content": {"title": "The Diagonalized Newton Algorithm for Nonnegative Matrix Factorization", "decision": "conferencePoster-iclr2013-confe...	[Review]: Overview: This paper proposes an element-wise (diagonal Hessian) Newton method to speed up convergence of the multiplicative update algorithm (MU) for NMF problems. Monotonic progress is guaranteed by an element-wise fall-back mechanism to MU. At a minimal computational overhead, this is shown to be effect...	anonymous reviewer 482c	null	null	{"id": "EW5mE9upmnWp1", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362382860000, "tmdate": 1362382860000, "ddate": null, "number": 3, "content": {"title": "review of The Diagonalized Newton Algorithm for Nonnegative Matrix Factorization", "review": "Overview:\r\n\r\nThis pa...	{ "criticism": 0, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 5, "praise": 3, "presentation_and_reporting": 1, "results_and_discussion": 2, "suggestion_and_solution": 2, "total": 7 }	2	1.431833	0.568167	2.016554	0.22067	0.016554	0	0	0.142857	0.714286	0.428571	0.142857	0.285714	0.285714	{ "criticism": 0, "example": 0, "importance_and_relevance": 0.14285714285714285, "materials_and_methods": 0.7142857142857143, "praise": 0.42857142857142855, "presentation_and_reporting": 0.14285714285714285, "results_and_discussion": 0.2857142857142857, "suggestion_and_solution": 0.2857142857142857 }	2	iclr2013	openreview	null
gGivgRWZsLgY0	Clustering Learning for Robotic Vision	We present the clustering learning technique applied to multi-layer feedforward deep neural networks. We show that this unsupervised learning technique can compute network filters with only a few minutes and a much reduced set of pa- rameters. The goal of this paper is to promote the technique for general-purpose robot...	Clustering Learning for Robotic Vision Eugenio Culurciello∗ Purdue University euge@purdue.edu Jordan Bates Purdue University jtbates@purdue.edu Aysegul Dundar Purdue University adundar@purdue.edu J.A. Perez-Carrasco University of Seville jperez2@us.es Clement Farabet New York University cfarabet@nyu.edu Abstract We pre...	Eugenio Culurciello, Jordan Bates, Aysegul Dundar, Jose Carrasco, Clement Farabet	Unknown	2,013	{"id": "gGivgRWZsLgY0", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358277300000, "tmdate": 1358277300000, "ddate": null, "number": 28, "content": {"title": "Clustering Learning for Robotic Vision", "decision": "conferencePoster-iclr2013-workshop", "abstract": "We present th...	[Review]: Dear reviewers, we have fixed all issues that you have reported in your kind review of the manuscript and uploaded a revision.	Eugenio Culurciello, Jordan Bates, Aysegul Dundar, Jose Carrasco, Clement Farabet	null	null	{"id": "PiVQP7pKuhiR5", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363392540000, "tmdate": 1363392540000, "ddate": null, "number": 4, "content": {"title": "", "review": "Dear reviewers, we have fixed all issues that you have reported in your kind review of the manuscript an...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 1 }	0	-2.27267	2.27267	0.001723	0.014941	0.001723	0	0	0	0	0	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	0	iclr2013	openreview	null
gGivgRWZsLgY0	Clustering Learning for Robotic Vision	We present the clustering learning technique applied to multi-layer feedforward deep neural networks. We show that this unsupervised learning technique can compute network filters with only a few minutes and a much reduced set of pa- rameters. The goal of this paper is to promote the technique for general-purpose robot...	Clustering Learning for Robotic Vision Eugenio Culurciello∗ Purdue University euge@purdue.edu Jordan Bates Purdue University jtbates@purdue.edu Aysegul Dundar Purdue University adundar@purdue.edu J.A. Perez-Carrasco University of Seville jperez2@us.es Clement Farabet New York University cfarabet@nyu.edu Abstract We pre...	Eugenio Culurciello, Jordan Bates, Aysegul Dundar, Jose Carrasco, Clement Farabet	Unknown	2,013	{"id": "gGivgRWZsLgY0", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358277300000, "tmdate": 1358277300000, "ddate": null, "number": 28, "content": {"title": "Clustering Learning for Robotic Vision", "decision": "conferencePoster-iclr2013-workshop", "abstract": "We present th...	[Review]: The paper presents an application of clustering-based feature learning ('CL') to image recognition tasks and tracking tasks for robotics. The basic system uses a clustering algorithm to train filters from small patches and then applies them convolutionally using a sum-abs-difference (instead of inner product...	anonymous reviewer 5eb5	null	null	{"id": "NL-vN6tmpZNMh", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362195960000, "tmdate": 1362195960000, "ddate": null, "number": 5, "content": {"title": "review of Clustering Learning for Robotic Vision", "review": "The paper presents an application of clustering-based fe...	{ "criticism": 6, "example": 0, "importance_and_relevance": 4, "materials_and_methods": 17, "praise": 3, "presentation_and_reporting": 0, "results_and_discussion": 6, "suggestion_and_solution": 3, "total": 22 }	1.772727	0.49435	1.278377	1.816561	0.436403	0.043834	0.272727	0	0.181818	0.772727	0.136364	0	0.272727	0.136364	{ "criticism": 0.2727272727272727, "example": 0, "importance_and_relevance": 0.18181818181818182, "materials_and_methods": 0.7727272727272727, "praise": 0.13636363636363635, "presentation_and_reporting": 0, "results_and_discussion": 0.2727272727272727, "suggestion_and_solution": 0.13636363636363635 }	1.772727	iclr2013	openreview	null
gGivgRWZsLgY0	Clustering Learning for Robotic Vision	We present the clustering learning technique applied to multi-layer feedforward deep neural networks. We show that this unsupervised learning technique can compute network filters with only a few minutes and a much reduced set of pa- rameters. The goal of this paper is to promote the technique for general-purpose robot...	Clustering Learning for Robotic Vision Eugenio Culurciello∗ Purdue University euge@purdue.edu Jordan Bates Purdue University jtbates@purdue.edu Aysegul Dundar Purdue University adundar@purdue.edu J.A. Perez-Carrasco University of Seville jperez2@us.es Clement Farabet New York University cfarabet@nyu.edu Abstract We pre...	Eugenio Culurciello, Jordan Bates, Aysegul Dundar, Jose Carrasco, Clement Farabet	Unknown	2,013	{"id": "gGivgRWZsLgY0", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358277300000, "tmdate": 1358277300000, "ddate": null, "number": 28, "content": {"title": "Clustering Learning for Robotic Vision", "decision": "conferencePoster-iclr2013-workshop", "abstract": "We present th...	[Review]: I am very sympathetic to the aims of the authors: Find simple, effective and fast deep networks to understand sensor data. The authors defer some of the more interesting bits to future work however: they note that sum-abs-diff should be much more efficient in silicon implementation then convolution style r...	anonymous reviewer d6ae	null	null	{"id": "KOmskcVuMBOLt", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362354720000, "tmdate": 1362354720000, "ddate": null, "number": 1, "content": {"title": "review of Clustering Learning for Robotic Vision", "review": "I am very sympathetic to the aims of the authors:\r\nF...	{ "criticism": 13, "example": 0, "importance_and_relevance": 8, "materials_and_methods": 15, "praise": 12, "presentation_and_reporting": 12, "results_and_discussion": 3, "suggestion_and_solution": 4, "total": 38 }	1.763158	-8.100861	9.864019	1.808995	0.451882	0.045837	0.342105	0	0.210526	0.394737	0.315789	0.315789	0.078947	0.105263	{ "criticism": 0.34210526315789475, "example": 0, "importance_and_relevance": 0.21052631578947367, "materials_and_methods": 0.39473684210526316, "praise": 0.3157894736842105, "presentation_and_reporting": 0.3157894736842105, "results_and_discussion": 0.07894736842105263, "suggestion_and_solution": 0.105...	1.763158	iclr2013	openreview	null
gGivgRWZsLgY0	Clustering Learning for Robotic Vision	We present the clustering learning technique applied to multi-layer feedforward deep neural networks. We show that this unsupervised learning technique can compute network filters with only a few minutes and a much reduced set of pa- rameters. The goal of this paper is to promote the technique for general-purpose robot...	Clustering Learning for Robotic Vision Eugenio Culurciello∗ Purdue University euge@purdue.edu Jordan Bates Purdue University jtbates@purdue.edu Aysegul Dundar Purdue University adundar@purdue.edu J.A. Perez-Carrasco University of Seville jperez2@us.es Clement Farabet New York University cfarabet@nyu.edu Abstract We pre...	Eugenio Culurciello, Jordan Bates, Aysegul Dundar, Jose Carrasco, Clement Farabet	Unknown	2,013	{"id": "gGivgRWZsLgY0", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358277300000, "tmdate": 1358277300000, "ddate": null, "number": 28, "content": {"title": "Clustering Learning for Robotic Vision", "decision": "conferencePoster-iclr2013-workshop", "abstract": "We present th...	[Review]: # Summary This paper compares two types of filtering operator (linear filtering vs. distance filtering) in convolutional neural networks for image processing. The paper evaluates two fairly arbitrarily-chosen architectures on the CIFAR-10 and SVHN image labeling tasks, and shows that neither of these arch...	anonymous reviewer d2a7	null	null	{"id": "DGTnGO8CnrcPN", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362366000000, "tmdate": 1362366000000, "ddate": null, "number": 3, "content": {"title": "review of Clustering Learning for Robotic Vision", "review": "# Summary\r\n\r\nThis paper compares two types of filter...	{ "criticism": 8, "example": 2, "importance_and_relevance": 2, "materials_and_methods": 19, "praise": 1, "presentation_and_reporting": 7, "results_and_discussion": 5, "suggestion_and_solution": 4, "total": 30 }	1.6	-2.961122	4.561122	1.650187	0.477785	0.050187	0.266667	0.066667	0.066667	0.633333	0.033333	0.233333	0.166667	0.133333	{ "criticism": 0.26666666666666666, "example": 0.06666666666666667, "importance_and_relevance": 0.06666666666666667, "materials_and_methods": 0.6333333333333333, "praise": 0.03333333333333333, "presentation_and_reporting": 0.23333333333333334, "results_and_discussion": 0.16666666666666666, "suggestion_a...	1.6	iclr2013	openreview	null
gGivgRWZsLgY0	Clustering Learning for Robotic Vision	We present the clustering learning technique applied to multi-layer feedforward deep neural networks. We show that this unsupervised learning technique can compute network filters with only a few minutes and a much reduced set of pa- rameters. The goal of this paper is to promote the technique for general-purpose robot...	Clustering Learning for Robotic Vision Eugenio Culurciello∗ Purdue University euge@purdue.edu Jordan Bates Purdue University jtbates@purdue.edu Aysegul Dundar Purdue University adundar@purdue.edu J.A. Perez-Carrasco University of Seville jperez2@us.es Clement Farabet New York University cfarabet@nyu.edu Abstract We pre...	Eugenio Culurciello, Jordan Bates, Aysegul Dundar, Jose Carrasco, Clement Farabet	Unknown	2,013	{"id": "gGivgRWZsLgY0", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358277300000, "tmdate": 1358277300000, "ddate": null, "number": 28, "content": {"title": "Clustering Learning for Robotic Vision", "decision": "conferencePoster-iclr2013-workshop", "abstract": "We present th...	[Review]: we accept the poster presentation, thank you for organizing this!	Eugenio Culurciello, Jordan Bates, Aysegul Dundar, Jose Carrasco, Clement Farabet	null	null	{"id": "-YucDnyrcVDfe", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1364401500000, "tmdate": 1364401500000, "ddate": null, "number": 2, "content": {"title": "", "review": "we accept the poster presentation, thank you for organizing this!"}, "forum": "gGivgRWZsLgY0", "referent...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 1 }	0	-2.27267	2.27267	0.000907	0.007866	0.000907	0	0	0	0	0	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	0	iclr2013	openreview	null
g6Jl6J3aMs6a7	Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in DeSTIN	This paper presents a basic enhancement to the DeSTIN deep learning architecture by replacing the explicitly calculated transition tables that are used to capture temporal features with a simpler, more scalable mechanism. This mechanism uses feedback of state information to cluster over a space comprised of both the sp...	Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in DeSTIN Steven R. Young and Itamar Arel Department of Electrical Engineering and Computer Science University of Tennessee {syoung22,itamar}@eecs.utk.edu Abstract This paper presents a basic enhancement to the DeSTIN deep learning architecture by repla...	Steven R. Young, Itamar Arel	Unknown	2,013	{"id": "g6Jl6J3aMs6a7", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358409600000, "tmdate": 1358409600000, "ddate": null, "number": 20, "content": {"decision": "reject", "title": "Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in\r\n DeSTIN", "abstract...	[Review]: The paper presents an extension to the author's prior 'DeSTIN' framework for spatio-temporal clustering. The lookup table that was previously used for state transitions is replaced by a feedback, output-to-input loop that somewhat resembles a recurrent neural network. However so little information is provided...	anonymous reviewer 675f	null	null	{"id": "GGdathbFl15ug", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362391440000, "tmdate": 1362391440000, "ddate": null, "number": 2, "content": {"title": "review of Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in\r\n DeSTIN", "review": "The paper p...	{ "criticism": 3, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 8, "praise": 0, "presentation_and_reporting": 2, "results_and_discussion": 3, "suggestion_and_solution": 2, "total": 12 }	1.5	1.484218	0.015782	1.562192	0.579202	0.062192	0.25	0	0	0.666667	0	0.166667	0.25	0.166667	{ "criticism": 0.25, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.6666666666666666, "praise": 0, "presentation_and_reporting": 0.16666666666666666, "results_and_discussion": 0.25, "suggestion_and_solution": 0.16666666666666666 }	1.5	iclr2013	openreview	null
g6Jl6J3aMs6a7	Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in DeSTIN	This paper presents a basic enhancement to the DeSTIN deep learning architecture by replacing the explicitly calculated transition tables that are used to capture temporal features with a simpler, more scalable mechanism. This mechanism uses feedback of state information to cluster over a space comprised of both the sp...	Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in DeSTIN Steven R. Young and Itamar Arel Department of Electrical Engineering and Computer Science University of Tennessee {syoung22,itamar}@eecs.utk.edu Abstract This paper presents a basic enhancement to the DeSTIN deep learning architecture by repla...	Steven R. Young, Itamar Arel	Unknown	2,013	{"id": "g6Jl6J3aMs6a7", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358409600000, "tmdate": 1358409600000, "ddate": null, "number": 20, "content": {"decision": "reject", "title": "Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in\r\n DeSTIN", "abstract...	[Review]: Improves the DeSTIN architecture by the same authors. They write on MNIST: A classification accuracy of 98.71% was achieved which is comparable to results using the first-generation DeSTIN architecture [1] and to results achieved with other state-of-the-art methods [4, 5, 6]. However, the error rate...	anonymous reviewer 6b68	null	null	{"id": "8BGL8F0WLpBcE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362163920000, "tmdate": 1362163920000, "ddate": null, "number": 1, "content": {"title": "review of Recurrent Online Clustering as a Spatio-Temporal Feature Extractor in\r\n DeSTIN", "review": "Improves th...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 2, "praise": 0, "presentation_and_reporting": 1, "results_and_discussion": 1, "suggestion_and_solution": 2, "total": 4 }	1.5	0.221623	1.278377	1.530909	0.357979	0.030909	0	0	0	0.5	0	0.25	0.25	0.5	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.5, "praise": 0, "presentation_and_reporting": 0.25, "results_and_discussion": 0.25, "suggestion_and_solution": 0.5 }	1.5	iclr2013	openreview	null
fm5jfAwPbOfP6	Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines	One conjecture in both deep learning and classical connectionist viewpoint is that the biological brain implements certain kinds of deep networks as its back-end. However, to our knowledge, a detailed correspondence has not yet been set up, which is important if we want to bridge between neuroscience and machine learni...	arXiv:1210.8442v3 [cs.AI] 27 Jan 2013 Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines Louis Yuanlong Shao Department of Computer Science & Engineering The Ohio State University shaoyu@cse.ohio-state.edu Abstract One conjecture in both deep learning and classical connecti onis...	Yuanlong Shao	Unknown	2,013	{"id": "fm5jfAwPbOfP6", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358316900000, "tmdate": 1358316900000, "ddate": null, "number": 63, "content": {"title": "Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines", "decision": "conferencePo...	[Review]: This paper proposes a scheme for utilizing LNP model neurons to perform inference in Boltzmann Machines. The contribution of the work is to map a Boltzmann Machine network onto a set of LNP model units and to demonstrate inference in this model. The idea of using neural spiking models to represent probabi...	anonymous reviewer 4490	null	null	{"id": "QQ1JEKYFTIQhj", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362262200000, "tmdate": 1362262200000, "ddate": null, "number": 2, "content": {"title": "review of Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines", "review": "This ...	{ "criticism": 9, "example": 1, "importance_and_relevance": 8, "materials_and_methods": 28, "praise": 4, "presentation_and_reporting": 5, "results_and_discussion": 7, "suggestion_and_solution": 8, "total": 54 }	1.296296	-25.233968	26.530264	1.32993	0.3416	0.033633	0.166667	0.018519	0.148148	0.518519	0.074074	0.092593	0.12963	0.148148	{ "criticism": 0.16666666666666666, "example": 0.018518518518518517, "importance_and_relevance": 0.14814814814814814, "materials_and_methods": 0.5185185185185185, "praise": 0.07407407407407407, "presentation_and_reporting": 0.09259259259259259, "results_and_discussion": 0.12962962962962962, "suggestion_...	1.296296	iclr2013	openreview	null
fm5jfAwPbOfP6	Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines	One conjecture in both deep learning and classical connectionist viewpoint is that the biological brain implements certain kinds of deep networks as its back-end. However, to our knowledge, a detailed correspondence has not yet been set up, which is important if we want to bridge between neuroscience and machine learni...	arXiv:1210.8442v3 [cs.AI] 27 Jan 2013 Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines Louis Yuanlong Shao Department of Computer Science & Engineering The Ohio State University shaoyu@cse.ohio-state.edu Abstract One conjecture in both deep learning and classical connecti onis...	Yuanlong Shao	Unknown	2,013	{"id": "fm5jfAwPbOfP6", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358316900000, "tmdate": 1358316900000, "ddate": null, "number": 63, "content": {"title": "Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines", "decision": "conferencePo...	[Review]: Thank you very much for the valuable reviews and references! I learned quite a lot from reading the suggested papers. --> For Reviewer caa8: - Regarding the question raised in the end of your review, I think a somewhat related question is why neurons use spikes and whether we shall follow that in our c...	Yuanlong Shao	null	null	{"id": "B4qSE6NM3ZEOV", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362383640000, "tmdate": 1362383640000, "ddate": null, "number": 4, "content": {"title": "", "review": "Thank you very much for the valuable reviews and references! I learned quite a lot from reading the sugg...	{ "criticism": 3, "example": 1, "importance_and_relevance": 4, "materials_and_methods": 51, "praise": 6, "presentation_and_reporting": 5, "results_and_discussion": 21, "suggestion_and_solution": 14, "total": 94 }	1.117021	-102.4315	103.548522	1.198523	0.744386	0.081502	0.031915	0.010638	0.042553	0.542553	0.06383	0.053191	0.223404	0.148936	{ "criticism": 0.031914893617021274, "example": 0.010638297872340425, "importance_and_relevance": 0.0425531914893617, "materials_and_methods": 0.5425531914893617, "praise": 0.06382978723404255, "presentation_and_reporting": 0.05319148936170213, "results_and_discussion": 0.22340425531914893, "suggestion_...	1.117021	iclr2013	openreview	null
fm5jfAwPbOfP6	Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines	One conjecture in both deep learning and classical connectionist viewpoint is that the biological brain implements certain kinds of deep networks as its back-end. However, to our knowledge, a detailed correspondence has not yet been set up, which is important if we want to bridge between neuroscience and machine learni...	arXiv:1210.8442v3 [cs.AI] 27 Jan 2013 Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines Louis Yuanlong Shao Department of Computer Science & Engineering The Ohio State University shaoyu@cse.ohio-state.edu Abstract One conjecture in both deep learning and classical connecti onis...	Yuanlong Shao	Unknown	2,013	{"id": "fm5jfAwPbOfP6", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358316900000, "tmdate": 1358316900000, "ddate": null, "number": 63, "content": {"title": "Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines", "decision": "conferencePo...	[Review]: This paper argues that inference in Boltzmann machines can be performed using neurons modelled according to the Linear Nonlinear-Poisson model. The LNP model is first presented, then one variant of inference procedure for Boltzmann machine is introduced and a section shows that LNP neurons can implement it. E...	anonymous reviewer ef61	null	null	{"id": "88txIZ2gY7lJh", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362392700000, "tmdate": 1362392700000, "ddate": null, "number": 3, "content": {"title": "review of Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines", "review": "This ...	{ "criticism": 3, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 5, "praise": 0, "presentation_and_reporting": 2, "results_and_discussion": 3, "suggestion_and_solution": 0, "total": 7 }	2	1.431833	0.568167	2.007073	0.095607	0.007073	0.428571	0	0.142857	0.714286	0	0.285714	0.428571	0	{ "criticism": 0.42857142857142855, "example": 0, "importance_and_relevance": 0.14285714285714285, "materials_and_methods": 0.7142857142857143, "praise": 0, "presentation_and_reporting": 0.2857142857142857, "results_and_discussion": 0.42857142857142855, "suggestion_and_solution": 0 }	2	iclr2013	openreview	null
fm5jfAwPbOfP6	Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines	One conjecture in both deep learning and classical connectionist viewpoint is that the biological brain implements certain kinds of deep networks as its back-end. However, to our knowledge, a detailed correspondence has not yet been set up, which is important if we want to bridge between neuroscience and machine learni...	arXiv:1210.8442v3 [cs.AI] 27 Jan 2013 Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines Louis Yuanlong Shao Department of Computer Science & Engineering The Ohio State University shaoyu@cse.ohio-state.edu Abstract One conjecture in both deep learning and classical connecti onis...	Yuanlong Shao	Unknown	2,013	{"id": "fm5jfAwPbOfP6", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358316900000, "tmdate": 1358316900000, "ddate": null, "number": 63, "content": {"title": "Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines", "decision": "conferencePo...	[Review]: The paper provides an explicit connection between the linear-nonlinear-poisson (LNP) model of biological neural networks and the Boltzmann machine. The author proposes a semi-stochastic inference procedure on Boltzmann machines, with some tweaks, that can be considered equivalent to the inference of an LNP mo...	anonymous reviewer caa8	null	null	{"id": "1JfiMxWFQy15Z", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361988540000, "tmdate": 1361988540000, "ddate": null, "number": 1, "content": {"title": "review of Linear-Nonlinear-Poisson Neuron Networks Perform Bayesian Inference On Boltzmann Machines", "review": "The p...	{ "criticism": 3, "example": 0, "importance_and_relevance": 2, "materials_and_methods": 13, "praise": 1, "presentation_and_reporting": 1, "results_and_discussion": 3, "suggestion_and_solution": 2, "total": 20 }	1.25	0.476661	0.773339	1.26792	0.192868	0.01792	0.15	0	0.1	0.65	0.05	0.05	0.15	0.1	{ "criticism": 0.15, "example": 0, "importance_and_relevance": 0.1, "materials_and_methods": 0.65, "praise": 0.05, "presentation_and_reporting": 0.05, "results_and_discussion": 0.15, "suggestion_and_solution": 0.1 }	1.25	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: I updated the paper according the reviewers' comments, and included results with a dual-tree implementation of t-SNE in the appendix. The updated paper should appear on Arxiv soon.	Laurens van der Maaten	null	null	{"id": "pA91py2CW8AQg", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362758580000, "tmdate": 1362758580000, "ddate": null, "number": 7, "content": {"title": "", "review": "I updated the paper according the reviewers' comments, and included results with a dual-tree implementat...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 1, "suggestion_and_solution": 2, "total": 2 }	1.5	-0.409674	1.909674	1.50274	0.188758	0.00274	0	0	0	0	0	0	0.5	1	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0.5, "suggestion_and_solution": 1 }	1.5	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: Thanks a bunch for these insightful reviews and for the useful pointers to related work (some of which I was not aware of)! In preliminary experiments, I compared locality-sensitive hashing and vantage-point trees in the initial nearest-neighbor (in the high-dimensional space). I found vantage-point trees ...	Laurens van der Maaten	null	null	{"id": "TTxAqxZdhgIV0", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362330660000, "tmdate": 1362330660000, "ddate": null, "number": 1, "content": {"title": "", "review": "Thanks a bunch for these insightful reviews and for the useful pointers to related work (some of which I...	{ "criticism": 5, "example": 0, "importance_and_relevance": 5, "materials_and_methods": 15, "praise": 4, "presentation_and_reporting": 2, "results_and_discussion": 5, "suggestion_and_solution": 7, "total": 21 }	2.047619	1.037544	1.010076	2.085424	0.422055	0.037805	0.238095	0	0.238095	0.714286	0.190476	0.095238	0.238095	0.333333	{ "criticism": 0.23809523809523808, "example": 0, "importance_and_relevance": 0.23809523809523808, "materials_and_methods": 0.7142857142857143, "praise": 0.19047619047619047, "presentation_and_reporting": 0.09523809523809523, "results_and_discussion": 0.23809523809523808, "suggestion_and_solution": 0.33...	2.047619	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: In typical applications of Barnes-Hut (like t-SNE), the force nearly vanishes in the far field, which allows for averaging those far-field forces without losing much accuracy. In algorithms that minimize, e.g., the squared error between two sets of pairwise distances, I guess you could do the opposite. The...	Laurens van der Maaten	null	null	{"id": "Hy8wy4X01CHmD", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363113120000, "tmdate": 1363113120000, "ddate": null, "number": 9, "content": {"title": "", "review": "In typical applications of Barnes-Hut (like t-SNE), the force nearly vanishes in the far field, which al...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 3, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 3, "suggestion_and_solution": 2, "total": 6 }	1.5	0.726661	0.773339	1.511222	0.162296	0.011222	0.166667	0	0	0.5	0	0	0.5	0.333333	{ "criticism": 0.16666666666666666, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.5, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0.5, "suggestion_and_solution": 0.3333333333333333 }	1.5	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: Great work, congratulations! It seems we and you have simultaneously found essentially the same solution. Our paper and software are here: Zhirong Yang, Jaakko Peltonen, Samuel Kaski. Scalable Optimization of Neighbor Embedding for Visualization. Accepted to ICML2013. Preprint and software: http://resea...	Zhirong Yang	null	null	{"id": "H3-iUVuyZzUgh", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1365114600000, "tmdate": 1365114600000, "ddate": null, "number": 6, "content": {"title": "", "review": "Great work, congratulations! It seems we and you have simultaneously found essentially the same solution...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 1, "praise": 1, "presentation_and_reporting": 0, "results_and_discussion": 1, "suggestion_and_solution": 0, "total": 6 }	0.5	-0.273339	0.773339	0.504121	0.040733	0.004121	0	0	0	0.166667	0.166667	0	0.166667	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.16666666666666666, "praise": 0.16666666666666666, "presentation_and_reporting": 0, "results_and_discussion": 0.16666666666666666, "suggestion_and_solution": 0 }	0.5	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: I have experimented with dual-tree variants of my algorithm (which required only trivial changes in the existing code), experimenting with both quadtrees and kd-trees as the underlying tree structures. Perhaps surprisingly, the dual-tree algorithm has approximately the same accuracy-speed trade-off as the Bar...	Laurens van der Maaten	null	null	{"id": "DyHSDHfKmbDPM", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362421080000, "tmdate": 1362421080000, "ddate": null, "number": 2, "content": {"title": "", "review": "I have experimented with dual-tree variants of my algorithm (which required only trivial changes in the ...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 5, "praise": 0, "presentation_and_reporting": 2, "results_and_discussion": 3, "suggestion_and_solution": 3, "total": 8 }	1.625	1.230439	0.394561	1.641607	0.216484	0.016607	0	0	0	0.625	0	0.25	0.375	0.375	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.625, "praise": 0, "presentation_and_reporting": 0.25, "results_and_discussion": 0.375, "suggestion_and_solution": 0.375 }	1.625	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: Stochastic neighbour embedding (SNE) is a sound, probabilistic method for dimensionality reduction. One of its limitations is that its complexity is O(N^2), where N is the, typically large, number of data points. To surmount this limitation, the this paper proposes computational methods to reduce the computa...	anonymous reviewer d9db	null	null	{"id": "Dkj3DFf4GZJPh", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362177000000, "tmdate": 1362177000000, "ddate": null, "number": 3, "content": {"title": "review of Barnes-Hut-SNE", "review": "Stochastic neighbour embedding (SNE) is a sound, probabilistic method for dimens...	{ "criticism": 5, "example": 5, "importance_and_relevance": 4, "materials_and_methods": 29, "praise": 7, "presentation_and_reporting": 5, "results_and_discussion": 10, "suggestion_and_solution": 12, "total": 68 }	1.132353	-46.609497	47.74185	1.202935	0.65562	0.070582	0.073529	0.073529	0.058824	0.426471	0.102941	0.073529	0.147059	0.176471	{ "criticism": 0.07352941176470588, "example": 0.07352941176470588, "importance_and_relevance": 0.058823529411764705, "materials_and_methods": 0.4264705882352941, "praise": 0.10294117647058823, "presentation_and_reporting": 0.07352941176470588, "results_and_discussion": 0.14705882352941177, "suggestion_...	1.132353	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: Laurens, have you thought about using similar ideas for embedding algorithms that also exploit global similarities (like multidimensional scaling)? I think in many types of data analysis, this can be extremely important.	Alex Bronstein	null	null	{"id": "AZcnMdQBqGZS4", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362833640000, "tmdate": 1362833640000, "ddate": null, "number": 5, "content": {"title": "", "review": "Laurens, have you thought about using similar ideas for embedding algorithms that also exploit global si...	{ "criticism": 0, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 2, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 2 }	1.5	-0.409674	1.909674	1.503515	0.095476	0.003515	0	0	0.5	1	0	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0.5, "materials_and_methods": 1, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	1.5	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: The submitted paper proposes a more efficient implementation of the Student-t distributed version of SNE. t-SNE is O(n^2), and the proposed implementation is O(nlogn). This offers a substantial improvement in the efficiency, such that very large datasets may be embedded. Furthermore, the speed increase is obt...	anonymous reviewer 7db1	null	null	{"id": "2VfI2cAZSF2P0", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362192420000, "tmdate": 1362192420000, "ddate": null, "number": 4, "content": {"title": "review of Barnes-Hut-SNE", "review": "The submitted paper proposes a more efficient implementation of the Student-t di...	{ "criticism": 0, "example": 0, "importance_and_relevance": 2, "materials_and_methods": 6, "praise": 5, "presentation_and_reporting": 4, "results_and_discussion": 3, "suggestion_and_solution": 0, "total": 14 }	1.428571	1.412789	0.015782	1.446719	0.185914	0.018148	0	0	0.142857	0.428571	0.357143	0.285714	0.214286	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0.14285714285714285, "materials_and_methods": 0.42857142857142855, "praise": 0.35714285714285715, "presentation_and_reporting": 0.2857142857142857, "results_and_discussion": 0.21428571428571427, "suggestion_and_solution": 0 }	1.428571	iclr2013	openreview	null
eQWJec0ursynH	Barnes-Hut-SNE	The paper presents an O(N log N)-implementation of t-SNE -- an embedding technique that is commonly used for the visualization of high-dimensional data in scatter plots and that normally runs in O(N^2). The new implementation uses vantage-point trees to compute sparse pairwise similarities between the input data object...	Barnes-Hut-SNE Laurens van der Maaten Pattern Recognition and Bioinformatics Group, Delft University of Technology Mekelweg 4, 2628 CD Delft, The Netherlands lvdmaaten@gmail.com Abstract The paper presents an O(Nlog N)-implementation of t-SNE — an embedding technique that is commonly used for the visualization of high-...	Laurens van der Maaten	Unknown	2,013	{"id": "eQWJec0ursynH", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358343900000, "tmdate": 1358343900000, "ddate": null, "number": 19, "content": {"title": "Barnes-Hut-SNE", "decision": "conferenceOral-iclr2013-conference", "abstract": "The paper presents an O(N log N)-impl...	[Review]: The paper addresses the problem of low-dimensional data embedding for visualization purposes via stochastic neighbor embedding, in which Euclidean dissimilarities in the data space are modulated by the Gaussian kernel, and a configuration of points in the low-dimensional embedding space is found such that the...	anonymous reviewer c262	null	null	{"id": "24bs4th0sfgwE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362833520000, "tmdate": 1362833520000, "ddate": null, "number": 8, "content": {"title": "review of Barnes-Hut-SNE", "review": "The paper addresses the problem of low-dimensional data embedding for visualizat...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 6, "praise": 3, "presentation_and_reporting": 0, "results_and_discussion": 2, "suggestion_and_solution": 1, "total": 15 }	0.866667	0.803537	0.06313	0.885755	0.184162	0.019088	0.066667	0	0	0.4	0.2	0	0.133333	0.066667	{ "criticism": 0.06666666666666667, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.4, "praise": 0.2, "presentation_and_reporting": 0, "results_and_discussion": 0.13333333333333333, "suggestion_and_solution": 0.06666666666666667 }	0.866667	iclr2013	openreview	null
bI58OFtQlLOQ7	Deep Learning for Detecting Robotic Grasps	In this work, we consider the problem of detecting robotic grasps in an RGB-D view of a scene containing objects. We present a two-step cascaded structure, where we have two deep networks, with the top detections from the first one re-evaluated by the second one. The first deep network has fewer features, is therefore ...	Deep Learning for Detecting Robotic Grasps Ian Lenz,† Honglak Lee,∗ and Ashutosh Saxena † † Department of Computer Science, Cornell University. ∗ Department of EECS, University of Michigan, Ann Arbor. Email: ianlenz@cs.cornell.edu, honglak@eecs.umich.edu, asaxena@cs.cornell.edu Abstract—We consider the problem of detec...	Ian Lenz, Honglak Lee, Ashutosh Saxena	Unknown	2,013	{"id": "bI58OFtQlLOQ7", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358403300000, "tmdate": 1358403300000, "ddate": null, "number": 61, "content": {"title": "Deep Learning for Detecting Robotic Grasps", "decision": "conferenceOral-iclr2013-workshop", "abstract": "In this wor...	[Review]: Summary: this paper uses the common 2-step procedure to first eliminate most of unlikely detection windows (high recall), then use a network with higher capacity for better discrimination (high precision). Deep learning (in the unsupervised sense) helps having features optimized for each of these 2 different ...	anonymous reviewer b096	null	null	{"id": "Sl9E4V1iE8lfU", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362192180000, "tmdate": 1362192180000, "ddate": null, "number": 2, "content": {"title": "review of Deep Learning for Detecting Robotic Grasps", "review": "Summary: this paper uses the common 2-step procedure...	{ "criticism": 1, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 10, "praise": 3, "presentation_and_reporting": 2, "results_and_discussion": 1, "suggestion_and_solution": 3, "total": 14 }	1.5	1.484218	0.015782	1.508858	0.131081	0.008858	0.071429	0	0.071429	0.714286	0.214286	0.142857	0.071429	0.214286	{ "criticism": 0.07142857142857142, "example": 0, "importance_and_relevance": 0.07142857142857142, "materials_and_methods": 0.7142857142857143, "praise": 0.21428571428571427, "presentation_and_reporting": 0.14285714285714285, "results_and_discussion": 0.07142857142857142, "suggestion_and_solution": 0.21...	1.5	iclr2013	openreview	null
bI58OFtQlLOQ7	Deep Learning for Detecting Robotic Grasps	In this work, we consider the problem of detecting robotic grasps in an RGB-D view of a scene containing objects. We present a two-step cascaded structure, where we have two deep networks, with the top detections from the first one re-evaluated by the second one. The first deep network has fewer features, is therefore ...	Deep Learning for Detecting Robotic Grasps Ian Lenz,† Honglak Lee,∗ and Ashutosh Saxena † † Department of Computer Science, Cornell University. ∗ Department of EECS, University of Michigan, Ann Arbor. Email: ianlenz@cs.cornell.edu, honglak@eecs.umich.edu, asaxena@cs.cornell.edu Abstract—We consider the problem of detec...	Ian Lenz, Honglak Lee, Ashutosh Saxena	Unknown	2,013	{"id": "bI58OFtQlLOQ7", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358403300000, "tmdate": 1358403300000, "ddate": null, "number": 61, "content": {"title": "Deep Learning for Detecting Robotic Grasps", "decision": "conferenceOral-iclr2013-workshop", "abstract": "In this wor...	[Review]: This paper uses a two-pass detection mechanism with sparse autoencoders for robotic grasp detection, a new application of deep learning. The methods used are fairly standard by now (two pass and autoencoders), so the main novelty of the paper is its nice application. It shows good results, which are well pres...	anonymous reviewer cf06	null	null	{"id": "Fsg-G38UWSlUP", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362414180000, "tmdate": 1362414180000, "ddate": null, "number": 1, "content": {"title": "review of Deep Learning for Detecting Robotic Grasps", "review": "This paper uses a two-pass detection mechanism with ...	{ "criticism": 1, "example": 0, "importance_and_relevance": 4, "materials_and_methods": 4, "praise": 3, "presentation_and_reporting": 1, "results_and_discussion": 1, "suggestion_and_solution": 1, "total": 5 }	3	1.989924	1.010076	3.004876	0.182278	0.004876	0.2	0	0.8	0.8	0.6	0.2	0.2	0.2	{ "criticism": 0.2, "example": 0, "importance_and_relevance": 0.8, "materials_and_methods": 0.8, "praise": 0.6, "presentation_and_reporting": 0.2, "results_and_discussion": 0.2, "suggestion_and_solution": 0.2 }	3	iclr2013	openreview	null
aQZtOGDyp-Ozh	Learning Stable Group Invariant Representations with Convolutional Networks	Transformation groups, such as translations or rotations, effectively express part of the variability observed in many recognition problems. The group structure enables the construction of invariant signal representations with appealing mathematical properties, where convolutions, together with pooling operators, bring...	arXiv:1301.3537v1 [cs.AI] 16 Jan 2013 Learning Stable Group Invariant Representations with Convolutional Networks Joan Bruna, Arthur Szlam and Y ann LeCun Courant Institute New York University New Nork, NY , 10013 {bruna,lecun}@cims.nyu.edu 1 Introduction Many signal categories in vision and auditory problems are inv...	Joan Bruna, Arthur Szlam, Yann LeCun	Unknown	2,013	{"id": "aQZtOGDyp-Ozh", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358488800000, "tmdate": 1358488800000, "ddate": null, "number": 62, "content": {"title": "Learning Stable Group Invariant Representations with Convolutional\r\n Networks", "decision": "conferencePoster-ic...	[Review]: I fully admit that I don't know enough about group theory to evaluate this submission. However, I do know about convolutional networks, so it is troubling that I can't understand it. Since this is only a workshop paper, we're not going to look for a new reviewer. When you do eventually pursue conference...	anonymous reviewer bf60	null	null	{"id": "uLsKzjPT0lx8V", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361928660000, "tmdate": 1361928660000, "ddate": null, "number": 3, "content": {"title": "review of Learning Stable Group Invariant Representations with Convolutional\r\n Networks", "review": "I fully admi...	{ "criticism": 4, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 1, "results_and_discussion": 0, "suggestion_and_solution": 2, "total": 6 }	1.333333	0.559994	0.773339	1.354211	0.261009	0.020878	0.666667	0	0.166667	0	0	0.166667	0	0.333333	{ "criticism": 0.6666666666666666, "example": 0, "importance_and_relevance": 0.16666666666666666, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0.16666666666666666, "results_and_discussion": 0, "suggestion_and_solution": 0.3333333333333333 }	1.333333	iclr2013	openreview	null
aQZtOGDyp-Ozh	Learning Stable Group Invariant Representations with Convolutional Networks	Transformation groups, such as translations or rotations, effectively express part of the variability observed in many recognition problems. The group structure enables the construction of invariant signal representations with appealing mathematical properties, where convolutions, together with pooling operators, bring...	arXiv:1301.3537v1 [cs.AI] 16 Jan 2013 Learning Stable Group Invariant Representations with Convolutional Networks Joan Bruna, Arthur Szlam and Y ann LeCun Courant Institute New York University New Nork, NY , 10013 {bruna,lecun}@cims.nyu.edu 1 Introduction Many signal categories in vision and auditory problems are inv...	Joan Bruna, Arthur Szlam, Yann LeCun	Unknown	2,013	{"id": "aQZtOGDyp-Ozh", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358488800000, "tmdate": 1358488800000, "ddate": null, "number": 62, "content": {"title": "Learning Stable Group Invariant Representations with Convolutional\r\n Networks", "decision": "conferencePoster-ic...	[Review]: This short paper presents a discussion on the nature and the type of invariances that are represented and learned by convolutional neural networks. It claims that the invariance a layer in a convolutional neural network can be expressed with a Lie group, and that the invariance of a deep convolutional neural...	anonymous reviewer 3316	null	null	{"id": "s1Kr1S64z0s8a", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362379800000, "tmdate": 1362379800000, "ddate": null, "number": 1, "content": {"title": "review of Learning Stable Group Invariant Representations with Convolutional\r\n Networks", "review": "This short p...	{ "criticism": 3, "example": 0, "importance_and_relevance": 4, "materials_and_methods": 2, "praise": 1, "presentation_and_reporting": 4, "results_and_discussion": 5, "suggestion_and_solution": 2, "total": 8 }	2.625	2.230439	0.394561	2.656455	0.386464	0.031455	0.375	0	0.5	0.25	0.125	0.5	0.625	0.25	{ "criticism": 0.375, "example": 0, "importance_and_relevance": 0.5, "materials_and_methods": 0.25, "praise": 0.125, "presentation_and_reporting": 0.5, "results_and_discussion": 0.625, "suggestion_and_solution": 0.25 }	2.625	iclr2013	openreview	null
aQZtOGDyp-Ozh	Learning Stable Group Invariant Representations with Convolutional Networks	Transformation groups, such as translations or rotations, effectively express part of the variability observed in many recognition problems. The group structure enables the construction of invariant signal representations with appealing mathematical properties, where convolutions, together with pooling operators, bring...	arXiv:1301.3537v1 [cs.AI] 16 Jan 2013 Learning Stable Group Invariant Representations with Convolutional Networks Joan Bruna, Arthur Szlam and Y ann LeCun Courant Institute New York University New Nork, NY , 10013 {bruna,lecun}@cims.nyu.edu 1 Introduction Many signal categories in vision and auditory problems are inv...	Joan Bruna, Arthur Szlam, Yann LeCun	Unknown	2,013	{"id": "aQZtOGDyp-Ozh", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358488800000, "tmdate": 1358488800000, "ddate": null, "number": 62, "content": {"title": "Learning Stable Group Invariant Representations with Convolutional\r\n Networks", "decision": "conferencePoster-ic...	[Review]: I would like to thank the reviewers for their time and constructive comments. Indeed, the paper, in its current form, explores the connection between deep convolutional networks and group invariance; but it lacks practical examples to motivate why this connection might be useful or interesting. I completely...	Joan Bruna	null	null	{"id": "7XaieIunN4X1I", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363658220000, "tmdate": 1363658220000, "ddate": null, "number": 2, "content": {"title": "", "review": "I would like to thank the reviewers for their time and constructive comments.\r\nIndeed, the paper, in i...	{ "criticism": 2, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 0, "praise": 2, "presentation_and_reporting": 1, "results_and_discussion": 0, "suggestion_and_solution": 2, "total": 5 }	1.6	0.589924	1.010076	1.613087	0.209465	0.013087	0.4	0	0.2	0	0.4	0.2	0	0.4	{ "criticism": 0.4, "example": 0, "importance_and_relevance": 0.2, "materials_and_methods": 0, "praise": 0.4, "presentation_and_reporting": 0.2, "results_and_discussion": 0, "suggestion_and_solution": 0.4 }	1.6	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: Authors propose an interesting idea to use deep neural networks with tied weights (recurrent architecture) for image classification. However, I am not familiar enough with the prior work to judge novelty of the idea. On the critical note, the paper is not easy to read without good knowledge of prior work, ...	anonymous reviewer a32e	null	null	{"id": "zzUEFMPkQcqkJ", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362400920000, "tmdate": 1362400920000, "ddate": null, "number": 4, "content": {"title": "review of Discriminative Recurrent Sparse Auto-Encoders", "review": "Authors propose an interesting idea to use deep n...	{ "criticism": 3, "example": 0, "importance_and_relevance": 3, "materials_and_methods": 2, "praise": 2, "presentation_and_reporting": 1, "results_and_discussion": 1, "suggestion_and_solution": 1, "total": 4 }	3.25	1.971623	1.278377	3.261614	0.245696	0.011614	0.75	0	0.75	0.5	0.5	0.25	0.25	0.25	{ "criticism": 0.75, "example": 0, "importance_and_relevance": 0.75, "materials_and_methods": 0.5, "praise": 0.5, "presentation_and_reporting": 0.25, "results_and_discussion": 0.25, "suggestion_and_solution": 0.25 }	3.25	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: Interesting implementation and results. But how is this approach related to the original, unmentioned work on Recurrent Auto-Encoders (RAAMs) by Pollack (1990) and colleagues? What's the main difference, if any? Similar for previous applications of RAAMs to unsupervised history compression, e.g., (Gisslen...	Jürgen Schmidhuber	null	null	{"id": "yy9FyB6XUYyiJ", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362604500000, "tmdate": 1362604500000, "ddate": null, "number": 10, "content": {"title": "", "review": "Interesting implementation and results. \r\n\r\nBut how is this approach related to the original, unmen...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 5, "praise": 1, "presentation_and_reporting": 1, "results_and_discussion": 4, "suggestion_and_solution": 1, "total": 10 }	1.2	1.057958	0.142042	1.209366	0.108205	0.009366	0	0	0	0.5	0.1	0.1	0.4	0.1	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.5, "praise": 0.1, "presentation_and_reporting": 0.1, "results_and_discussion": 0.4, "suggestion_and_solution": 0.1 }	1.2	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: Minor side comment: IN GENERAL, having a cost term at each iteration (time step of the unfolded network) does not eliminate the vanishing gradient problem!!! The short-term dependencies can now be learned through the gradient on the cost on the early iterations, but the long-term effects may still be impro...	Yann LeCun	null	null	{"id": "vCQPfwXgPoCu7", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1364571960000, "tmdate": 1364571960000, "ddate": null, "number": 3, "content": {"title": "", "review": "Minor side comment: IN GENERAL, having a cost term at each iteration (time step of the unfolded network)...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 5, "praise": 2, "presentation_and_reporting": 0, "results_and_discussion": 4, "suggestion_and_solution": 1, "total": 7 }	1.714286	1.146118	0.568167	1.725235	0.133498	0.010949	0	0	0	0.714286	0.285714	0	0.571429	0.142857	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.7142857142857143, "praise": 0.2857142857142857, "presentation_and_reporting": 0, "results_and_discussion": 0.5714285714285714, "suggestion_and_solution": 0.14285714285714285 }	1.714286	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: SUMMARY: The authors describe a discriminative recurrent sparse auto-encoder, which is essentially a recurrent neural network with a fixed input and linear rectifier units. The auto-encoder is initially trained to reproduce digits of MNIST, while enforcing a sparse representation. In a later phase it is tr...	anonymous reviewer bc93	null	null	{"id": "uc38pbD6RhB1Z", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363316520000, "tmdate": 1363316520000, "ddate": null, "number": 7, "content": {"title": "review of Discriminative Recurrent Sparse Auto-Encoders", "review": "SUMMARY:\r\n\r\nThe authors describe a discrimina...	{ "criticism": 1, "example": 5, "importance_and_relevance": 5, "materials_and_methods": 17, "praise": 6, "presentation_and_reporting": 12, "results_and_discussion": 7, "suggestion_and_solution": 9, "total": 38 }	1.631579	-8.23244	9.864019	1.678906	0.475321	0.047327	0.026316	0.131579	0.131579	0.447368	0.157895	0.315789	0.184211	0.236842	{ "criticism": 0.02631578947368421, "example": 0.13157894736842105, "importance_and_relevance": 0.13157894736842105, "materials_and_methods": 0.4473684210526316, "praise": 0.15789473684210525, "presentation_and_reporting": 0.3157894736842105, "results_and_discussion": 0.18421052631578946, "suggestion_an...	1.631579	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: Thank you for this interesting contribution. The differentiation of hidden units into class units and parts units is fascinating and connects with what I consider a central objective for deep learning, i.e., learning representations where the learned features disentangle the underlying factors of variation (a...	Yoshua Bengio	null	null	{"id": "__De_0xQMv_R3", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361907180000, "tmdate": 1361907180000, "ddate": null, "number": 11, "content": {"title": "", "review": "Thank you for this interesting contribution. The differentiation of hidden units into class units and p...	{ "criticism": 0, "example": 1, "importance_and_relevance": 2, "materials_and_methods": 7, "praise": 3, "presentation_and_reporting": 2, "results_and_discussion": 4, "suggestion_and_solution": 1, "total": 13 }	1.538462	1.538462	0	1.559149	0.221667	0.020687	0	0.076923	0.153846	0.538462	0.230769	0.153846	0.307692	0.076923	{ "criticism": 0, "example": 0.07692307692307693, "importance_and_relevance": 0.15384615384615385, "materials_and_methods": 0.5384615384615384, "praise": 0.23076923076923078, "presentation_and_reporting": 0.15384615384615385, "results_and_discussion": 0.3076923076923077, "suggestion_and_solution": 0.076...	1.538462	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: * Jurgen Schidhuber: Thank you very much for your constructive comments. 1. Like the work of Pollack (1990), DrSAE is based on an recursive autoencoder that receives input on each iteration. However, (sequential) RAAMs iteratively add new information on each iteration, and then iteratively reconstruct...	Jason Rolfe	null	null	{"id": "UEx3pAOcLlpPT", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363223340000, "tmdate": 1363223340000, "ddate": null, "number": 9, "content": {"title": "", "review": "* Jurgen Schidhuber:\r\n\r\nThank you very much for your constructive comments.\r\n\r\n1. Like the work...	{ "criticism": 0, "example": 1, "importance_and_relevance": 0, "materials_and_methods": 9, "praise": 0, "presentation_and_reporting": 3, "results_and_discussion": 1, "suggestion_and_solution": 4, "total": 18 }	1	0.605439	0.394561	1.023958	0.251051	0.023958	0	0.055556	0	0.5	0	0.166667	0.055556	0.222222	{ "criticism": 0, "example": 0.05555555555555555, "importance_and_relevance": 0, "materials_and_methods": 0.5, "praise": 0, "presentation_and_reporting": 0.16666666666666666, "results_and_discussion": 0.05555555555555555, "suggestion_and_solution": 0.2222222222222222 }	1	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: Hi, This looks a whole lot like the semi-supervised recursive autoencoder that we introduced at EMNLP 2011 [1] and the unfolding recursive autoencoder that we introduced at NIPS 2011. These models also have a reconstruction + cross entropy error at every iteration and hence do not suffer from the vanish...	Richard Socher	null	null	{"id": "TTDqPocbXWPbU", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1364548920000, "tmdate": 1364548920000, "ddate": null, "number": 12, "content": {"title": "", "review": "Hi,\r\n\r\nThis looks a whole lot like the semi-supervised recursive autoencoder that we introduced at ...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 2, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 1, "suggestion_and_solution": 0, "total": 5 }	0.6	-0.410076	1.010076	0.605543	0.054058	0.005543	0	0	0	0.4	0	0	0.2	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.4, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0.2, "suggestion_and_solution": 0 }	0.6	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: Summary and general overview: ---------------------------------------------- The paper introduces Discriminative Recurrent Sparse Auto-Encoders, a new model, but more importantly a careful analysis of the behaviour of this model. It suggests that the hidden layers of the model learn to differentiate into a ...	anonymous reviewer 8ddb	null	null	{"id": "SKcvK2UDvgKxL", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362177060000, "tmdate": 1362177060000, "ddate": null, "number": 1, "content": {"title": "review of Discriminative Recurrent Sparse Auto-Encoders", "review": "Summary and general overview:\r\n----------------...	{ "criticism": 3, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 17, "praise": 1, "presentation_and_reporting": 5, "results_and_discussion": 5, "suggestion_and_solution": 6, "total": 36 }	1.055556	-7.29335	8.348905	1.089782	0.335076	0.034226	0.083333	0	0.027778	0.472222	0.027778	0.138889	0.138889	0.166667	{ "criticism": 0.08333333333333333, "example": 0, "importance_and_relevance": 0.027777777777777776, "materials_and_methods": 0.4722222222222222, "praise": 0.027777777777777776, "presentation_and_reporting": 0.1388888888888889, "results_and_discussion": 0.1388888888888889, "suggestion_and_solution": 0.16...	1.055556	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: Interesting work! The use of relU units in an RNN is something I haven't seen before. I'd be interested in some discussion on how relU compares to e.g. tanh units in the recurrent setting. I imagine relU units may suffer less from vanishing/saturation during RNN training. We have a related model (deep dis...	Andrew Maas	null	null	{"id": "PZqMVyiGDoPcE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363734420000, "tmdate": 1363734420000, "ddate": null, "number": 6, "content": {"title": "", "review": "Interesting work! The use of relU units in an RNN is something I haven't seen before. I'd be interested ...	{ "criticism": 0, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 4, "praise": 1, "presentation_and_reporting": 1, "results_and_discussion": 2, "suggestion_and_solution": 2, "total": 11 }	1	0.93687	0.06313	1.007182	0.108629	0.007182	0	0	0.090909	0.363636	0.090909	0.090909	0.181818	0.181818	{ "criticism": 0, "example": 0, "importance_and_relevance": 0.09090909090909091, "materials_and_methods": 0.36363636363636365, "praise": 0.09090909090909091, "presentation_and_reporting": 0.09090909090909091, "results_and_discussion": 0.18181818181818182, "suggestion_and_solution": 0.18181818181818182 }	1	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: We are very thankful to all the reviewers and commenters for their constructive comments. * Anonymous 8ddb: 1. Indeed, the architecture of DrSAE is similar to a deep sparse rectifier neural network (Glorot, Bordes, and Bengio, 2011) with tied weights (Bengio, Boulanger-Lewandowski and Pascanu, 2012). I...	Jason Rolfe	null	null	{"id": "NNXtqijEtiN98", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363222920000, "tmdate": 1363222920000, "ddate": null, "number": 2, "content": {"title": "", "review": "We are very thankful to all the reviewers and commenters for their constructive comments.\r\n\r\n* Anony...	{ "criticism": 5, "example": 2, "importance_and_relevance": 7, "materials_and_methods": 29, "praise": 4, "presentation_and_reporting": 10, "results_and_discussion": 11, "suggestion_and_solution": 9, "total": 60 }	1.283333	-33.580054	34.863387	1.349696	0.622362	0.066362	0.083333	0.033333	0.116667	0.483333	0.066667	0.166667	0.183333	0.15	{ "criticism": 0.08333333333333333, "example": 0.03333333333333333, "importance_and_relevance": 0.11666666666666667, "materials_and_methods": 0.48333333333333334, "praise": 0.06666666666666667, "presentation_and_reporting": 0.16666666666666666, "results_and_discussion": 0.18333333333333332, "suggestion_...	1.283333	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: The paper describes the following variation of an autoencoder: An encoder (with relu nonlinearity) is iterated for 11 steps, with observations providing biases for the hiddens at each step. Afterwards, a decoder reconstructs the data from the last-step hiddens. In addition, a softmax computes class-labels fro...	anonymous reviewer dd6a	null	null	{"id": "4V-Ozm5k8mVcn", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363400280000, "tmdate": 1363400280000, "ddate": null, "number": 8, "content": {"title": "review of Discriminative Recurrent Sparse Auto-Encoders", "review": "The paper describes the following variation of an...	{ "criticism": 3, "example": 6, "importance_and_relevance": 3, "materials_and_methods": 20, "praise": 5, "presentation_and_reporting": 9, "results_and_discussion": 6, "suggestion_and_solution": 7, "total": 33 }	1.787879	-4.525093	6.312972	1.820305	0.339924	0.032426	0.090909	0.181818	0.090909	0.606061	0.151515	0.272727	0.181818	0.212121	{ "criticism": 0.09090909090909091, "example": 0.18181818181818182, "importance_and_relevance": 0.09090909090909091, "materials_and_methods": 0.6060606060606061, "praise": 0.15151515151515152, "presentation_and_reporting": 0.2727272727272727, "results_and_discussion": 0.18181818181818182, "suggestion_an...	1.787879	iclr2013	openreview	null
aJh-lFL2dFJ21	Discriminative Recurrent Sparse Auto-Encoders	We present the discriminative recurrent sparse auto-encoder model, which consists of an encoder whose hidden layer is recurrent, and two linear decoders, one to reconstruct the input, and one to predict the output. The hidden layer is composed of rectified linear units (ReLU) and is subject to a sparsity penalty. The n...	Discriminative Recurrent Sparse Auto-Encoders Jason Tyler Rolfe & Yann LeCun Courant Institute of Mathematical Sciences, New York University 719 Broadway, 12th Floor New York, NY 10003 {rolfe, yann}@cs.nyu.edu Abstract We present the discriminative recurrent sparse auto-encoder model, comprising a recurrent encoder of ...	Jason Rolfe, Yann LeCun	Unknown	2,013	{"id": "aJh-lFL2dFJ21", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358412300000, "tmdate": 1358412300000, "ddate": null, "number": 51, "content": {"title": "Discriminative Recurrent Sparse Auto-Encoders", "decision": "conferenceOral-iclr2013-conference", "abstract": "We pre...	[Review]: Hi Jason and Yann, Thanks for the insightful reply. Best, Richard	Richard Socher	null	null	{"id": "-uMO-UhKgU-Z_", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1368275760000, "tmdate": 1368275760000, "ddate": null, "number": 5, "content": {"title": "", "review": "Hi Jason and Yann,\r\nThanks for the insightful reply.\r\nBest,\r\nRichard"}, "forum": "aJh-lFL2dFJ21", ...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 1, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 2 }	0.5	-1.409674	1.909674	0.500793	0.011873	0.000793	0	0	0	0	0.5	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0.5, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	0.5	iclr2013	openreview	null
ZhGJ9KQlXi9jk	Complexity of Representation and Inference in Compositional Models with Part Sharing	This paper describes serial and parallel compositional models of multiple objects with part sharing. Objects are built by part-subpart compositions and expressed in terms of a hierarchical dictionary of object parts. These parts are represented on lattices of decreasing sizes which yield an executive summary descriptio...	Complexity of Representation and Inference in Compositional Models with Part Sharing Alan L. Yuille Depts. of Statistics, Computer Science & Psychology University of California, Los Angeles yuille@stat.ucla.edu Roozbeh Mottaghi Department of Computer Science University of California, Los Angeles roozbehm@cs.ucla.edu Ab...	Alan Yuille, Roozbeh Mottaghi	Unknown	2,013	{"id": "ZhGJ9KQlXi9jk", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358405100000, "tmdate": 1358405100000, "ddate": null, "number": 34, "content": {"title": "Complexity of Representation and Inference in Compositional Models with\r\n Part Sharing", "decision": "conference...	[Review]: Reviewer c1e8, Please read the authors' responses to your review. Do they change your evaluation of the paper?	Aaron Courville	null	null	{"id": "sPw_squDz1sCV", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363536060000, "tmdate": 1363536060000, "ddate": null, "number": 4, "content": {"title": "", "review": "Reviewer c1e8,\r\n\r\nPlease read the authors' responses to your review. Do they change your evaluation...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 1, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 1, "total": 2 }	1	-0.909674	1.909674	1.001189	0.095306	0.001189	0	0	0	0.5	0	0	0	0.5	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.5, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0.5 }	1	iclr2013	openreview	null
ZhGJ9KQlXi9jk	Complexity of Representation and Inference in Compositional Models with Part Sharing	This paper describes serial and parallel compositional models of multiple objects with part sharing. Objects are built by part-subpart compositions and expressed in terms of a hierarchical dictionary of object parts. These parts are represented on lattices of decreasing sizes which yield an executive summary descriptio...	Complexity of Representation and Inference in Compositional Models with Part Sharing Alan L. Yuille Depts. of Statistics, Computer Science & Psychology University of California, Los Angeles yuille@stat.ucla.edu Roozbeh Mottaghi Department of Computer Science University of California, Los Angeles roozbehm@cs.ucla.edu Ab...	Alan Yuille, Roozbeh Mottaghi	Unknown	2,013	{"id": "ZhGJ9KQlXi9jk", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358405100000, "tmdate": 1358405100000, "ddate": null, "number": 34, "content": {"title": "Complexity of Representation and Inference in Compositional Models with\r\n Part Sharing", "decision": "conference...	[Review]: The paper describe a compositional object models that take the form of a hierarchical generative models. Both object and part models provide (1) a set of part models, and (2) a generative model essentially describing how parts are composed. A distinctive feature of this model is the ability to support 'part ...	anonymous reviewer c1e8	null	null	{"id": "p7BE8U1NHl8Tr", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361997540000, "tmdate": 1361997540000, "ddate": null, "number": 2, "content": {"title": "review of Complexity of Representation and Inference in Compositional Models with\r\n Part Sharing", "review": "The...	{ "criticism": 0, "example": 0, "importance_and_relevance": 3, "materials_and_methods": 6, "praise": 0, "presentation_and_reporting": 1, "results_and_discussion": 1, "suggestion_and_solution": 0, "total": 9 }	1.222222	0.969703	0.252519	1.235192	0.158	0.012969	0	0	0.333333	0.666667	0	0.111111	0.111111	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0.3333333333333333, "materials_and_methods": 0.6666666666666666, "praise": 0, "presentation_and_reporting": 0.1111111111111111, "results_and_discussion": 0.1111111111111111, "suggestion_and_solution": 0 }	1.222222	iclr2013	openreview	null
ZhGJ9KQlXi9jk	Complexity of Representation and Inference in Compositional Models with Part Sharing	This paper describes serial and parallel compositional models of multiple objects with part sharing. Objects are built by part-subpart compositions and expressed in terms of a hierarchical dictionary of object parts. These parts are represented on lattices of decreasing sizes which yield an executive summary descriptio...	Complexity of Representation and Inference in Compositional Models with Part Sharing Alan L. Yuille Depts. of Statistics, Computer Science & Psychology University of California, Los Angeles yuille@stat.ucla.edu Roozbeh Mottaghi Department of Computer Science University of California, Los Angeles roozbehm@cs.ucla.edu Ab...	Alan Yuille, Roozbeh Mottaghi	Unknown	2,013	{"id": "ZhGJ9KQlXi9jk", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358405100000, "tmdate": 1358405100000, "ddate": null, "number": 34, "content": {"title": "Complexity of Representation and Inference in Compositional Models with\r\n Part Sharing", "decision": "conference...	[Review]: This paper presents a complexity analysis of certain inference algorithms for compositional models of images based on part sharing. The intuition behind these models is that objects are composed of parts and that each of these parts can appear in many different objects; with sensible parallels (not mentio...	anonymous reviewer 915e	null	null	{"id": "oCzZPts6ZYo6d", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362211680000, "tmdate": 1362211680000, "ddate": null, "number": 1, "content": {"title": "review of Complexity of Representation and Inference in Compositional Models with\r\n Part Sharing", "review": "Thi...	{ "criticism": 1, "example": 0, "importance_and_relevance": 2, "materials_and_methods": 3, "praise": 3, "presentation_and_reporting": 0, "results_and_discussion": 1, "suggestion_and_solution": 0, "total": 5 }	2	0.989924	1.010076	2.008548	0.134112	0.008548	0.2	0	0.4	0.6	0.6	0	0.2	0	{ "criticism": 0.2, "example": 0, "importance_and_relevance": 0.4, "materials_and_methods": 0.6, "praise": 0.6, "presentation_and_reporting": 0, "results_and_discussion": 0.2, "suggestion_and_solution": 0 }	2	iclr2013	openreview	null
ZhGJ9KQlXi9jk	Complexity of Representation and Inference in Compositional Models with Part Sharing	This paper describes serial and parallel compositional models of multiple objects with part sharing. Objects are built by part-subpart compositions and expressed in terms of a hierarchical dictionary of object parts. These parts are represented on lattices of decreasing sizes which yield an executive summary descriptio...	Complexity of Representation and Inference in Compositional Models with Part Sharing Alan L. Yuille Depts. of Statistics, Computer Science & Psychology University of California, Los Angeles yuille@stat.ucla.edu Roozbeh Mottaghi Department of Computer Science University of California, Los Angeles roozbehm@cs.ucla.edu Ab...	Alan Yuille, Roozbeh Mottaghi	Unknown	2,013	{"id": "ZhGJ9KQlXi9jk", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358405100000, "tmdate": 1358405100000, "ddate": null, "number": 34, "content": {"title": "Complexity of Representation and Inference in Compositional Models with\r\n Part Sharing", "decision": "conference...	[Review]: This paper explores how inference can be done in a part-sharing model and the computational cost of doing so. It relies on 'executive summaries' where each layer only holds approximate information about the layer below. The authors also study the computational complexity of this inference in various settings....	anonymous reviewer a9e8	null	null	{"id": "EHF-pZ3qwbnAT", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362609900000, "tmdate": 1362609900000, "ddate": null, "number": 3, "content": {"title": "review of Complexity of Representation and Inference in Compositional Models with\r\n Part Sharing", "review": "Thi...	{ "criticism": 1, "example": 0, "importance_and_relevance": 3, "materials_and_methods": 4, "praise": 1, "presentation_and_reporting": 1, "results_and_discussion": 3, "suggestion_and_solution": 1, "total": 8 }	1.75	1.355439	0.394561	1.760347	0.163458	0.010347	0.125	0	0.375	0.5	0.125	0.125	0.375	0.125	{ "criticism": 0.125, "example": 0, "importance_and_relevance": 0.375, "materials_and_methods": 0.5, "praise": 0.125, "presentation_and_reporting": 0.125, "results_and_discussion": 0.375, "suggestion_and_solution": 0.125 }	1.75	iclr2013	openreview	null
YBi6KFA7PfKo5	Two SVDs produce more focal deep learning representations	A key characteristic of work on deep learning and neural networks in general is that it relies on representations of the input that support generalization, robust inference, domain adaptation and other desirable functionalities. Much recent progress in the field has focused on efficient and effective methods for comput...	arXiv:1301.3627v2 [cs.CL] 11 May 2013 Two SVDs produce more focal deep learning representations Hinrich Sch¨ utze Center for Information and Language Processing University of Munich, Germany hs999@ifnlp.org Christian Scheible Institute for NLP University of Stuttgart, Germany scheibcn@ims.uni-stuttgart.de Abstract A ...	Hinrich Schuetze, Christian Scheible	Unknown	2,013	{"id": "YBi6KFA7PfKo5", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358457300000, "tmdate": 1358457300000, "ddate": null, "number": 42, "content": {"title": "Two SVDs produce more focal deep learning representations", "decision": "conferencePoster-iclr2013-workshop", "abstra...	[Review]: Thanks for your comments! The suggestions seem all good and pertinent to us and (in case the paper should be accepted and assuming there is enough space) we will incorporate them when revising the paper. In particular: relate the new method to overview in Turney&Pantel, to kernel PCA and matrix factorization ...	Hinrich Schuetze	null	null	{"id": "aK4z5qBF7bEod", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363717680000, "tmdate": 1363717680000, "ddate": null, "number": 3, "content": {"title": "", "review": "Thanks for your comments! The suggestions seem all good and pertinent to us and (in case the paper shoul...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 4, "praise": 1, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 3, "total": 6 }	1.5	0.726661	0.773339	1.524102	0.298963	0.024102	0.166667	0	0	0.666667	0.166667	0	0	0.5	{ "criticism": 0.16666666666666666, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.6666666666666666, "praise": 0.16666666666666666, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0.5 }	1.5	iclr2013	openreview	null
YBi6KFA7PfKo5	Two SVDs produce more focal deep learning representations	A key characteristic of work on deep learning and neural networks in general is that it relies on representations of the input that support generalization, robust inference, domain adaptation and other desirable functionalities. Much recent progress in the field has focused on efficient and effective methods for comput...	arXiv:1301.3627v2 [cs.CL] 11 May 2013 Two SVDs produce more focal deep learning representations Hinrich Sch¨ utze Center for Information and Language Processing University of Munich, Germany hs999@ifnlp.org Christian Scheible Institute for NLP University of Stuttgart, Germany scheibcn@ims.uni-stuttgart.de Abstract A ...	Hinrich Schuetze, Christian Scheible	Unknown	2,013	{"id": "YBi6KFA7PfKo5", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358457300000, "tmdate": 1358457300000, "ddate": null, "number": 42, "content": {"title": "Two SVDs produce more focal deep learning representations", "decision": "conferencePoster-iclr2013-workshop", "abstra...	[Review]: This paper proposes to use two consecutive SVDs to produce a continuous representation. This paper also introduces a property called focality. They claim that this property may be important for neural network: many classifiers cannot efficiently handle conjunctions of several features unless they are expl...	anonymous reviewer 2448	null	null	{"id": "VFwT2CLWfA2kU", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361986620000, "tmdate": 1361986620000, "ddate": null, "number": 1, "content": {"title": "review of Two SVDs produce more focal deep learning representations", "review": "This paper proposes to use two consec...	{ "criticism": 2, "example": 1, "importance_and_relevance": 2, "materials_and_methods": 4, "praise": 1, "presentation_and_reporting": 5, "results_and_discussion": 4, "suggestion_and_solution": 4, "total": 12 }	1.916667	1.900884	0.015782	1.944917	0.330766	0.028251	0.166667	0.083333	0.166667	0.333333	0.083333	0.416667	0.333333	0.333333	{ "criticism": 0.16666666666666666, "example": 0.08333333333333333, "importance_and_relevance": 0.16666666666666666, "materials_and_methods": 0.3333333333333333, "praise": 0.08333333333333333, "presentation_and_reporting": 0.4166666666666667, "results_and_discussion": 0.3333333333333333, "suggestion_and...	1.916667	iclr2013	openreview	null
YBi6KFA7PfKo5	Two SVDs produce more focal deep learning representations	A key characteristic of work on deep learning and neural networks in general is that it relies on representations of the input that support generalization, robust inference, domain adaptation and other desirable functionalities. Much recent progress in the field has focused on efficient and effective methods for comput...	arXiv:1301.3627v2 [cs.CL] 11 May 2013 Two SVDs produce more focal deep learning representations Hinrich Sch¨ utze Center for Information and Language Processing University of Munich, Germany hs999@ifnlp.org Christian Scheible Institute for NLP University of Stuttgart, Germany scheibcn@ims.uni-stuttgart.de Abstract A ...	Hinrich Schuetze, Christian Scheible	Unknown	2,013	{"id": "YBi6KFA7PfKo5", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358457300000, "tmdate": 1358457300000, "ddate": null, "number": 42, "content": {"title": "Two SVDs produce more focal deep learning representations", "decision": "conferencePoster-iclr2013-workshop", "abstra...	[Review]: This paper introduces a novel method to induce word vector representations from a corpus of unlabeled text. The method relies upon 'stacking' singular value decomposition with an intermediate normalization nonlinearity. The authors propose 'focality' as a metric for quantifying the quality of a learned repres...	anonymous reviewer 4c9d	null	null	{"id": "3wTuUWS9F_w4i", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362188640000, "tmdate": 1362188640000, "ddate": null, "number": 2, "content": {"title": "review of Two SVDs produce more focal deep learning representations", "review": "This paper introduces a novel method ...	{ "criticism": 1, "example": 0, "importance_and_relevance": 3, "materials_and_methods": 22, "praise": 5, "presentation_and_reporting": 11, "results_and_discussion": 2, "suggestion_and_solution": 10, "total": 24 }	2.25	0.340326	1.909674	2.310243	0.619804	0.060243	0.041667	0	0.125	0.916667	0.208333	0.458333	0.083333	0.416667	{ "criticism": 0.041666666666666664, "example": 0, "importance_and_relevance": 0.125, "materials_and_methods": 0.9166666666666666, "praise": 0.20833333333333334, "presentation_and_reporting": 0.4583333333333333, "results_and_discussion": 0.08333333333333333, "suggestion_and_solution": 0.4166666666666667...	2.25	iclr2013	openreview	null
V_-8VUqv8h_H3	The Manifold of Human Emotions	Sentiment analysis predicts the presence of positive or negative emotions in a text document. In this paper, we consider higher dimensional extensions of the sentiment concept, which represent a richer set of human emotions. Our approach goes beyond previous work in that our model contains a continuous manifold rather ...	arXiv:1301.3214v1 [cs.CL] 15 Jan 2013 The Manifold of Human Emotions Seungyeon Kim, Fuxin Li, Guy Lebanon, Irfan Essa College of Computing Georgia Institute of Technology {seungyeon.kim@, fli@cc., lebanon@cc., irfan@cc.}gatech.edu Abstract Sentiment analysis predicts the presence of positive or negative emotions in a...	Seungyeon Kim, Fuxin Li, Guy Lebanon, Irfan Essa	Unknown	2,013	{"id": "V_-8VUqv8h_H3", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358332200000, "tmdate": 1358332200000, "ddate": null, "number": 31, "content": {"title": "The Manifold of Human Emotions", "decision": "conferencePoster-iclr2013-workshop", "abstract": "Sentiment analysis pr...	[Review]: This paper proposes a new method for sentiment analysis of text documents based on two phases: first, learning a continuous vector representation of the document (a projection on the mood manifold) and second, learning to map from this representation to the sentiment classes. The assumption behind this mo...	anonymous reviewer e0d0	null	null	{"id": "C4MuPqjpEwP7S", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362239340000, "tmdate": 1362239340000, "ddate": null, "number": 1, "content": {"title": "review of The Manifold of Human Emotions", "review": "This paper proposes a new method for sentiment analysis of text\...	{ "criticism": 4, "example": 1, "importance_and_relevance": 1, "materials_and_methods": 7, "praise": 2, "presentation_and_reporting": 5, "results_and_discussion": 1, "suggestion_and_solution": 2, "total": 11 }	2.090909	2.027779	0.06313	2.167749	0.723474	0.07684	0.363636	0.090909	0.090909	0.636364	0.181818	0.454545	0.090909	0.181818	{ "criticism": 0.36363636363636365, "example": 0.09090909090909091, "importance_and_relevance": 0.09090909090909091, "materials_and_methods": 0.6363636363636364, "praise": 0.18181818181818182, "presentation_and_reporting": 0.45454545454545453, "results_and_discussion": 0.09090909090909091, "suggestion_a...	2.090909	iclr2013	openreview	null
V_-8VUqv8h_H3	The Manifold of Human Emotions	Sentiment analysis predicts the presence of positive or negative emotions in a text document. In this paper, we consider higher dimensional extensions of the sentiment concept, which represent a richer set of human emotions. Our approach goes beyond previous work in that our model contains a continuous manifold rather ...	arXiv:1301.3214v1 [cs.CL] 15 Jan 2013 The Manifold of Human Emotions Seungyeon Kim, Fuxin Li, Guy Lebanon, Irfan Essa College of Computing Georgia Institute of Technology {seungyeon.kim@, fli@cc., lebanon@cc., irfan@cc.}gatech.edu Abstract Sentiment analysis predicts the presence of positive or negative emotions in a...	Seungyeon Kim, Fuxin Li, Guy Lebanon, Irfan Essa	Unknown	2,013	{"id": "V_-8VUqv8h_H3", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358332200000, "tmdate": 1358332200000, "ddate": null, "number": 31, "content": {"title": "The Manifold of Human Emotions", "decision": "conferencePoster-iclr2013-workshop", "abstract": "Sentiment analysis pr...	[Review]: This paper introduces a model for sentiment analysis aimed at capturing blended, non-binary notions of sentiment. The paper uses a novel dataset of >1 million blog posts (livejournal) using 32 emoticons as labels. The model uses a Gaussian latent variable to embed bag of words documents into a vector space sh...	anonymous reviewer 9992	null	null	{"id": "ADj5N2hoX0_ox", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362105540000, "tmdate": 1362105540000, "ddate": null, "number": 2, "content": {"title": "review of The Manifold of Human Emotions", "review": "This paper introduces a model for sentiment analysis aimed at ca...	{ "criticism": 7, "example": 0, "importance_and_relevance": 4, "materials_and_methods": 10, "praise": 2, "presentation_and_reporting": 5, "results_and_discussion": 5, "suggestion_and_solution": 3, "total": 17 }	2.117647	1.865128	0.252519	2.18739	0.675849	0.069743	0.411765	0	0.235294	0.588235	0.117647	0.294118	0.294118	0.176471	{ "criticism": 0.4117647058823529, "example": 0, "importance_and_relevance": 0.23529411764705882, "materials_and_methods": 0.5882352941176471, "praise": 0.11764705882352941, "presentation_and_reporting": 0.29411764705882354, "results_and_discussion": 0.29411764705882354, "suggestion_and_solution": 0.176...	2.117647	iclr2013	openreview	null
UUwuUaQ5qRyWn	When Does a Mixture of Products Contain a Product of Mixtures?	We prove results on the relative representational power of mixtures of product distributions and restricted Boltzmann machines (products of mixtures of pairs of product distributions). Tools of independent interest are mode-based polyhedral approximations sensitive enough to compare full-dimensional models, and charact...	When Does a Mixture of Products Contain a Product of Mixtures? Guido F. Mont´ ufar∗1,2 and Jason Morton†2 1Max Planck Institute for Mathematics in the Sciences, Inselstrasse 22, 04103 Leipzig, Germany. 2Department of Mathematics, Pennsylvania State University, University Park, PA 16802, USA. September 22, 2014 Abstract...	Guido F. Montufar, Jason Morton	Unknown	2,013	{"id": "UUwuUaQ5qRyWn", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1357861500000, "tmdate": 1357861500000, "ddate": null, "number": 10, "content": {"title": "When Does a Mixture of Products Contain a Product of Mixtures?", "decision": "conferencePoster-iclr2013-workshop", "a...	[Review]: The paper analyses the representational capacity of RBM's, contrasting it with other simple models. I think the results are new but I'm definitely not an expert on this field. They are likely to be interesting for people working on RBM's, and thus to people at ICLR.	anonymous reviewer 91ea	null	null	{"id": "dYGvTnylo5TlF", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361559180000, "tmdate": 1361559180000, "ddate": null, "number": 4, "content": {"title": "review of When Does a Mixture of Products Contain a Product of Mixtures?", "review": "The paper analyses the represent...	{ "criticism": 1, "example": 0, "importance_and_relevance": 2, "materials_and_methods": 1, "praise": 1, "presentation_and_reporting": 0, "results_and_discussion": 1, "suggestion_and_solution": 0, "total": 3 }	2	0.421757	1.578243	2.001474	0.099448	0.001474	0.333333	0	0.666667	0.333333	0.333333	0	0.333333	0	{ "criticism": 0.3333333333333333, "example": 0, "importance_and_relevance": 0.6666666666666666, "materials_and_methods": 0.3333333333333333, "praise": 0.3333333333333333, "presentation_and_reporting": 0, "results_and_discussion": 0.3333333333333333, "suggestion_and_solution": 0 }	2	iclr2013	openreview	null
UUwuUaQ5qRyWn	When Does a Mixture of Products Contain a Product of Mixtures?	We prove results on the relative representational power of mixtures of product distributions and restricted Boltzmann machines (products of mixtures of pairs of product distributions). Tools of independent interest are mode-based polyhedral approximations sensitive enough to compare full-dimensional models, and charact...	When Does a Mixture of Products Contain a Product of Mixtures? Guido F. Mont´ ufar∗1,2 and Jason Morton†2 1Max Planck Institute for Mathematics in the Sciences, Inselstrasse 22, 04103 Leipzig, Germany. 2Department of Mathematics, Pennsylvania State University, University Park, PA 16802, USA. September 22, 2014 Abstract...	Guido F. Montufar, Jason Morton	Unknown	2,013	{"id": "UUwuUaQ5qRyWn", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1357861500000, "tmdate": 1357861500000, "ddate": null, "number": 10, "content": {"title": "When Does a Mixture of Products Contain a Product of Mixtures?", "decision": "conferencePoster-iclr2013-workshop", "a...	[Review]: This paper compares the representational power of Restricted Boltzmann Machines (RBMs) with that of mixtures of product distributions. The main result is that RBMs can be exponentially more efficient (in terms of the number of parameters required) to represent some classes of probability distributions. Thi...	anonymous reviewer 6c04	null	null	{"id": "dPNqPnWus1JhM", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362219240000, "tmdate": 1362219240000, "ddate": null, "number": 3, "content": {"title": "review of When Does a Mixture of Products Contain a Product of Mixtures?", "review": "This paper compares the represen...	{ "criticism": 4, "example": 7, "importance_and_relevance": 6, "materials_and_methods": 5, "praise": 4, "presentation_and_reporting": 7, "results_and_discussion": 8, "suggestion_and_solution": 5, "total": 22 }	2.090909	0.812532	1.278377	2.107778	0.228527	0.016869	0.181818	0.318182	0.272727	0.227273	0.181818	0.318182	0.363636	0.227273	{ "criticism": 0.18181818181818182, "example": 0.3181818181818182, "importance_and_relevance": 0.2727272727272727, "materials_and_methods": 0.22727272727272727, "praise": 0.18181818181818182, "presentation_and_reporting": 0.3181818181818182, "results_and_discussion": 0.36363636363636365, "suggestion_and...	2.090909	iclr2013	openreview	null
UUwuUaQ5qRyWn	When Does a Mixture of Products Contain a Product of Mixtures?	We prove results on the relative representational power of mixtures of product distributions and restricted Boltzmann machines (products of mixtures of pairs of product distributions). Tools of independent interest are mode-based polyhedral approximations sensitive enough to compare full-dimensional models, and charact...	When Does a Mixture of Products Contain a Product of Mixtures? Guido F. Mont´ ufar∗1,2 and Jason Morton†2 1Max Planck Institute for Mathematics in the Sciences, Inselstrasse 22, 04103 Leipzig, Germany. 2Department of Mathematics, Pennsylvania State University, University Park, PA 16802, USA. September 22, 2014 Abstract...	Guido F. Montufar, Jason Morton	Unknown	2,013	{"id": "UUwuUaQ5qRyWn", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1357861500000, "tmdate": 1357861500000, "ddate": null, "number": 10, "content": {"title": "When Does a Mixture of Products Contain a Product of Mixtures?", "decision": "conferencePoster-iclr2013-workshop", "a...	[Review]: This paper attempts at comparing mixture of factorial distributions (called product distributions) to RBMs. It does so by analyzing several theoretical properties, such as the smallest models which can represent any distribution with a given number of strong modes (or at least one of these distributions) or t...	anonymous reviewer 51ff	null	null	{"id": "boGLoNdiUmbgV", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362582360000, "tmdate": 1362582360000, "ddate": null, "number": 1, "content": {"title": "review of When Does a Mixture of Products Contain a Product of Mixtures?", "review": "This paper attempts at comparing...	{ "criticism": 6, "example": 2, "importance_and_relevance": 2, "materials_and_methods": 6, "praise": 0, "presentation_and_reporting": 8, "results_and_discussion": 7, "suggestion_and_solution": 5, "total": 15 }	2.4	2.33687	0.06313	2.411317	0.18545	0.011317	0.4	0.133333	0.133333	0.4	0	0.533333	0.466667	0.333333	{ "criticism": 0.4, "example": 0.13333333333333333, "importance_and_relevance": 0.13333333333333333, "materials_and_methods": 0.4, "praise": 0, "presentation_and_reporting": 0.5333333333333333, "results_and_discussion": 0.4666666666666667, "suggestion_and_solution": 0.3333333333333333 }	2.4	iclr2013	openreview	null
UUwuUaQ5qRyWn	When Does a Mixture of Products Contain a Product of Mixtures?	We prove results on the relative representational power of mixtures of product distributions and restricted Boltzmann machines (products of mixtures of pairs of product distributions). Tools of independent interest are mode-based polyhedral approximations sensitive enough to compare full-dimensional models, and charact...	When Does a Mixture of Products Contain a Product of Mixtures? Guido F. Mont´ ufar∗1,2 and Jason Morton†2 1Max Planck Institute for Mathematics in the Sciences, Inselstrasse 22, 04103 Leipzig, Germany. 2Department of Mathematics, Pennsylvania State University, University Park, PA 16802, USA. September 22, 2014 Abstract...	Guido F. Montufar, Jason Morton	Unknown	2,013	{"id": "UUwuUaQ5qRyWn", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1357861500000, "tmdate": 1357861500000, "ddate": null, "number": 10, "content": {"title": "When Does a Mixture of Products Contain a Product of Mixtures?", "decision": "conferencePoster-iclr2013-workshop", "a...	[Review]: We thank all three reviewers for the helpful comments, which enabled us to improve the paper. We have uploaded a revision to the arxiv taking into account the comments, and respond to some specific concerns below. We were unsure as to whether we should make the paper longer by providing more in-line intui...	Guido F. Montufar, Jason Morton	null	null	{"id": "FdwnFIZNOxF5S", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363384620000, "tmdate": 1363384620000, "ddate": null, "number": 2, "content": {"title": "", "review": "We thank all three reviewers for the helpful comments, which enabled us to improve the paper. We have u...	{ "criticism": 16, "example": 15, "importance_and_relevance": 2, "materials_and_methods": 18, "praise": 3, "presentation_and_reporting": 35, "results_and_discussion": 17, "suggestion_and_solution": 20, "total": 64 }	1.96875	-39.08135	41.0501	2.009593	0.423794	0.040843	0.25	0.234375	0.03125	0.28125	0.046875	0.546875	0.265625	0.3125	{ "criticism": 0.25, "example": 0.234375, "importance_and_relevance": 0.03125, "materials_and_methods": 0.28125, "praise": 0.046875, "presentation_and_reporting": 0.546875, "results_and_discussion": 0.265625, "suggestion_and_solution": 0.3125 }	1.96875	iclr2013	openreview	null
TT0bFo9VZpFWg	Big Neural Networks Waste Capacity	This article exposes the failure of some big neural networks to leverage added capacity to reduce underfitting. Past research suggest diminishing returns when increasing the size of neural networks. Our experiments on ImageNet LSVRC-2010 show that this may be due to the fact that bigger networks underfit the training o...	Big Neural Networks Waste Capacity Yann N. Dauphin & Yoshua Bengio D´epartement d’informatique et de recherche op´erationnelle Universit´e de Montr´eal, Montr´eal, QC, Canada Abstract This article exposes the failure of some big neural networks to leverage added capacity to reduce underﬁtting. Past research suggest dim...	Yann Dauphin, Yoshua Bengio	Unknown	2,013	{"id": "TT0bFo9VZpFWg", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358405100000, "tmdate": 1358405100000, "ddate": null, "number": 46, "content": {"title": "Big Neural Networks Waste Capacity", "decision": "conferenceOral-iclr2013-workshop", "abstract": "This article expose...	[Review]: The authors speculate that the inability of additional units to reduce the training error beyond a certain point in their experiments might be because 'networks with more capacity have more local minima.' How can this claim about local minima be reconciled with theoretical asymptotic results that show tha...	George Dahl	null	null	{"id": "PPZdA2YqSgAq6", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362402480000, "tmdate": 1362402480000, "ddate": null, "number": 4, "content": {"title": "", "review": "The authors speculate that the inability of additional units to reduce\r\nthe training error beyond a ce...	{ "criticism": 2, "example": 1, "importance_and_relevance": 0, "materials_and_methods": 10, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 11, "suggestion_and_solution": 1, "total": 14 }	1.785714	1.769932	0.015782	1.853917	0.61989	0.068203	0.142857	0.071429	0	0.714286	0	0	0.785714	0.071429	{ "criticism": 0.14285714285714285, "example": 0.07142857142857142, "importance_and_relevance": 0, "materials_and_methods": 0.7142857142857143, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0.7857142857142857, "suggestion_and_solution": 0.07142857142857142 }	1.785714	iclr2013	openreview	null
TT0bFo9VZpFWg	Big Neural Networks Waste Capacity	This article exposes the failure of some big neural networks to leverage added capacity to reduce underfitting. Past research suggest diminishing returns when increasing the size of neural networks. Our experiments on ImageNet LSVRC-2010 show that this may be due to the fact that bigger networks underfit the training o...	Big Neural Networks Waste Capacity Yann N. Dauphin & Yoshua Bengio D´epartement d’informatique et de recherche op´erationnelle Universit´e de Montr´eal, Montr´eal, QC, Canada Abstract This article exposes the failure of some big neural networks to leverage added capacity to reduce underﬁtting. Past research suggest dim...	Yann Dauphin, Yoshua Bengio	Unknown	2,013	{"id": "TT0bFo9VZpFWg", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358405100000, "tmdate": 1358405100000, "ddate": null, "number": 46, "content": {"title": "Big Neural Networks Waste Capacity", "decision": "conferenceOral-iclr2013-workshop", "abstract": "This article expose...	[Review]: The net gets bigger, yet keeps underfitting the training set. Authors suspect that gradient descent is the culprit. An interesting study!	anonymous reviewer b2da	null	null	{"id": "MvRrJo2NhwMOE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362019740000, "tmdate": 1362019740000, "ddate": null, "number": 3, "content": {"title": "review of Big Neural Networks Waste Capacity", "review": "The net gets bigger, yet keeps underfitting the training set...	{ "criticism": 0, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 0, "praise": 1, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 3 }	0.666667	-0.911576	1.578243	0.67145	0.081471	0.004783	0	0	0.333333	0	0.333333	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0.3333333333333333, "materials_and_methods": 0, "praise": 0.3333333333333333, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	0.666667	iclr2013	openreview	null
TT0bFo9VZpFWg	Big Neural Networks Waste Capacity	This article exposes the failure of some big neural networks to leverage added capacity to reduce underfitting. Past research suggest diminishing returns when increasing the size of neural networks. Our experiments on ImageNet LSVRC-2010 show that this may be due to the fact that bigger networks underfit the training o...	Big Neural Networks Waste Capacity Yann N. Dauphin & Yoshua Bengio D´epartement d’informatique et de recherche op´erationnelle Universit´e de Montr´eal, Montr´eal, QC, Canada Abstract This article exposes the failure of some big neural networks to leverage added capacity to reduce underﬁtting. Past research suggest dim...	Yann Dauphin, Yoshua Bengio	Unknown	2,013	{"id": "TT0bFo9VZpFWg", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358405100000, "tmdate": 1358405100000, "ddate": null, "number": 46, "content": {"title": "Big Neural Networks Waste Capacity", "decision": "conferenceOral-iclr2013-workshop", "abstract": "This article expose...	[Review]: This papers show the effects of under-fitting in a neural network as the size of a single neural network layer increases. The overall model is composed of SIFT extraction, k-mean, and this single hidden layer neural network. The paper suggest that this under-fitting problem is due to optimization problems wit...	anonymous reviewer 9741	null	null	{"id": "ChpzCSZ9zqCTR", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361967300000, "tmdate": 1361967300000, "ddate": null, "number": 2, "content": {"title": "review of Big Neural Networks Waste Capacity", "review": "This papers show the effects of under-fitting in a neural ne...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 6, "praise": 0, "presentation_and_reporting": 1, "results_and_discussion": 1, "suggestion_and_solution": 1, "total": 9 }	1.111111	0.858592	0.252519	1.147145	0.340193	0.036034	0.111111	0	0	0.666667	0	0.111111	0.111111	0.111111	{ "criticism": 0.1111111111111111, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.6666666666666666, "praise": 0, "presentation_and_reporting": 0.1111111111111111, "results_and_discussion": 0.1111111111111111, "suggestion_and_solution": 0.1111111111111111 }	1.111111	iclr2013	openreview	null
TT0bFo9VZpFWg	Big Neural Networks Waste Capacity	This article exposes the failure of some big neural networks to leverage added capacity to reduce underfitting. Past research suggest diminishing returns when increasing the size of neural networks. Our experiments on ImageNet LSVRC-2010 show that this may be due to the fact that bigger networks underfit the training o...	Big Neural Networks Waste Capacity Yann N. Dauphin & Yoshua Bengio D´epartement d’informatique et de recherche op´erationnelle Universit´e de Montr´eal, Montr´eal, QC, Canada Abstract This article exposes the failure of some big neural networks to leverage added capacity to reduce underﬁtting. Past research suggest dim...	Yann Dauphin, Yoshua Bengio	Unknown	2,013	{"id": "TT0bFo9VZpFWg", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358405100000, "tmdate": 1358405100000, "ddate": null, "number": 46, "content": {"title": "Big Neural Networks Waste Capacity", "decision": "conferenceOral-iclr2013-workshop", "abstract": "This article expose...	[Review]: Interesting topic. Another potential explanation for the diminishing return is the already good performance of networks with 5k hidden units. It could be that last bit of training performance requires fitting an especially difficult / nonlinear function and thus even 15k units in a single layer MLP can't do i...	Andrew Maas	null	null	{"id": "5w24FePB4ywro", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362373200000, "tmdate": 1362373200000, "ddate": null, "number": 1, "content": {"title": "", "review": "Interesting topic. Another potential explanation for the diminishing return is the already good performa...	{ "criticism": 1, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 4, "praise": 2, "presentation_and_reporting": 1, "results_and_discussion": 3, "suggestion_and_solution": 2, "total": 6 }	2.333333	1.559994	0.773339	2.35603	0.286777	0.022696	0.166667	0	0.166667	0.666667	0.333333	0.166667	0.5	0.333333	{ "criticism": 0.16666666666666666, "example": 0, "importance_and_relevance": 0.16666666666666666, "materials_and_methods": 0.6666666666666666, "praise": 0.3333333333333333, "presentation_and_reporting": 0.16666666666666666, "results_and_discussion": 0.5, "suggestion_and_solution": 0.3333333333333333 }	2.333333	iclr2013	openreview	null
SqNvxV9FQoSk2	Switched linear encoding with rectified linear autoencoders	Several recent results in machine learning have established formal connections between autoencoders---artificial neural network models that attempt to reproduce their inputs---and other coding models like sparse coding and K-means. This paper explores in depth an autoencoder model that is constructed using rectified li...	Switched linear coding with rectiﬁed linear autoencoders Leif Johnson Craig Corcoran Computer Science Department The University of Texas at Austin Austin, TX 78701 {leif,ccor}@cs.utexas.edu Abstract Several recent results in machine learning have established formal connections be- tween autoencoders—artiﬁcial neural ne...	Leif Johnson, Craig Corcoran	Unknown	2,013	{"id": "SqNvxV9FQoSk2", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358404200000, "tmdate": 1358404200000, "ddate": null, "number": 3, "content": {"decision": "reject", "title": "Switched linear encoding with rectified linear autoencoders", "abstract": "Several recent result...	[Review]: The paper draws links between autoencoders with tied weights and rectified linear units (similar to Glorot et al AISTATS 2011), the triangle k-means and soft-thresholding of Coates et al. (AISTATS 2011 and ICML 2011), and the linear-autoencoder-like ICA learning criterion of Le et al (NIPS 2011). The first ...	anonymous reviewer ab3b	null	null	{"id": "oozAQe0eAnQ1w", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362360840000, "tmdate": 1362360840000, "ddate": null, "number": 2, "content": {"title": "review of Switched linear encoding with rectified linear autoencoders", "review": "The paper draws links between autoe...	{ "criticism": 6, "example": 0, "importance_and_relevance": 3, "materials_and_methods": 10, "praise": 1, "presentation_and_reporting": 2, "results_and_discussion": 4, "suggestion_and_solution": 1, "total": 17 }	1.588235	1.335716	0.252519	1.627927	0.386483	0.039692	0.352941	0	0.176471	0.588235	0.058824	0.117647	0.235294	0.058824	{ "criticism": 0.35294117647058826, "example": 0, "importance_and_relevance": 0.17647058823529413, "materials_and_methods": 0.5882352941176471, "praise": 0.058823529411764705, "presentation_and_reporting": 0.11764705882352941, "results_and_discussion": 0.23529411764705882, "suggestion_and_solution": 0.0...	1.588235	iclr2013	openreview	null
SqNvxV9FQoSk2	Switched linear encoding with rectified linear autoencoders	Several recent results in machine learning have established formal connections between autoencoders---artificial neural network models that attempt to reproduce their inputs---and other coding models like sparse coding and K-means. This paper explores in depth an autoencoder model that is constructed using rectified li...	Switched linear coding with rectiﬁed linear autoencoders Leif Johnson Craig Corcoran Computer Science Department The University of Texas at Austin Austin, TX 78701 {leif,ccor}@cs.utexas.edu Abstract Several recent results in machine learning have established formal connections be- tween autoencoders—artiﬁcial neural ne...	Leif Johnson, Craig Corcoran	Unknown	2,013	{"id": "SqNvxV9FQoSk2", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358404200000, "tmdate": 1358404200000, "ddate": null, "number": 3, "content": {"decision": "reject", "title": "Switched linear encoding with rectified linear autoencoders", "abstract": "Several recent result...	[Review]: This paper analyzes properties of rectified linear autoencoder networks. In particular, the paper shows that rectified linear networks are similar to linear networks (ICA). The major difference is the nolinearity ('switching') that allows the decoder to select a subset of features. Such selection can ...	anonymous reviewer 9c3f	null	null	{"id": "kH1XHWcuGjDuU", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361946600000, "tmdate": 1361946600000, "ddate": null, "number": 3, "content": {"title": "review of Switched linear encoding with rectified linear autoencoders", "review": "This paper analyzes properties of r...	{ "criticism": 6, "example": 0, "importance_and_relevance": 2, "materials_and_methods": 12, "praise": 2, "presentation_and_reporting": 2, "results_and_discussion": 7, "suggestion_and_solution": 4, "total": 21 }	1.666667	0.656591	1.010076	1.699129	0.336207	0.032462	0.285714	0	0.095238	0.571429	0.095238	0.095238	0.333333	0.190476	{ "criticism": 0.2857142857142857, "example": 0, "importance_and_relevance": 0.09523809523809523, "materials_and_methods": 0.5714285714285714, "praise": 0.09523809523809523, "presentation_and_reporting": 0.09523809523809523, "results_and_discussion": 0.3333333333333333, "suggestion_and_solution": 0.1904...	1.666667	iclr2013	openreview	null
SqNvxV9FQoSk2	Switched linear encoding with rectified linear autoencoders	Several recent results in machine learning have established formal connections between autoencoders---artificial neural network models that attempt to reproduce their inputs---and other coding models like sparse coding and K-means. This paper explores in depth an autoencoder model that is constructed using rectified li...	Switched linear coding with rectiﬁed linear autoencoders Leif Johnson Craig Corcoran Computer Science Department The University of Texas at Austin Austin, TX 78701 {leif,ccor}@cs.utexas.edu Abstract Several recent results in machine learning have established formal connections be- tween autoencoders—artiﬁcial neural ne...	Leif Johnson, Craig Corcoran	Unknown	2,013	{"id": "SqNvxV9FQoSk2", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358404200000, "tmdate": 1358404200000, "ddate": null, "number": 3, "content": {"decision": "reject", "title": "Switched linear encoding with rectified linear autoencoders", "abstract": "Several recent result...	[Review]: In the deep learning community there has been a recent trend in moving away from the traditional sigmoid/tanh activation function to inject non-linearity into the model. One activation function that has been shown to work well in a number of cases is called Rectified Linear Unit (ReLU). Building on ...	anonymous reviewer 5a78	null	null	{"id": "ff2dqJ6VEpR8u", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362252900000, "tmdate": 1362252900000, "ddate": null, "number": 1, "content": {"title": "review of Switched linear encoding with rectified linear autoencoders", "review": "In the deep learning community ther...	{ "criticism": 8, "example": 1, "importance_and_relevance": 5, "materials_and_methods": 13, "praise": 0, "presentation_and_reporting": 2, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 18 }	1.611111	1.21655	0.394561	1.652421	0.402042	0.041309	0.444444	0.055556	0.277778	0.722222	0	0.111111	0	0	{ "criticism": 0.4444444444444444, "example": 0.05555555555555555, "importance_and_relevance": 0.2777777777777778, "materials_and_methods": 0.7222222222222222, "praise": 0, "presentation_and_reporting": 0.1111111111111111, "results_and_discussion": 0, "suggestion_and_solution": 0 }	1.611111	iclr2013	openreview	null
SSnY462CYz1Cu	Knowledge Matters: Importance of Prior Information for Optimization	We explore the effect of introducing prior information into the intermediate level of neural networks for a learning task on which all the state-of-the-art machine learning algorithms tested failed to learn. We motivate our work from the hypothesis that humans learn such intermediate concepts from other individuals via...	Knowledge Matters: Importance of Prior Information for Optimization C ¸ a˘ glar G¨ ul¸ cehre gulcehrc@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ erationnelle Universit´ e de Montr´ eal, Montr´ eal, QC, Canada Yoshua Bengio bengioy@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ ...	Çağlar Gülçehre, Yoshua Bengio	Unknown	2,013	{"id": "SSnY462CYz1Cu", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358528400000, "tmdate": 1358528400000, "ddate": null, "number": 36, "content": {"title": "Knowledge Matters: Importance of Prior Information for Optimization", "decision": "conferenceOral-iclr2013-conference...	[Review]: I would like to add some further comments for the purpose of constructive discussion. The authors try to provide further insights into why and when deep learning works, and to broaden the focus of the kind of questions usually asked in this community, in particular by making connections to biological cogni...	David Reichert	null	null	{"id": "nMIynqm1yCndY", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362784440000, "tmdate": 1362784440000, "ddate": null, "number": 3, "content": {"title": "", "review": "I would like to add some further comments for the purpose of constructive discussion.\r\n\r\nThe authors...	{ "criticism": 8, "example": 1, "importance_and_relevance": 6, "materials_and_methods": 12, "praise": 3, "presentation_and_reporting": 0, "results_and_discussion": 5, "suggestion_and_solution": 6, "total": 26 }	1.576923	-1.090308	2.667231	1.599828	0.272048	0.022905	0.307692	0.038462	0.230769	0.461538	0.115385	0	0.192308	0.230769	{ "criticism": 0.3076923076923077, "example": 0.038461538461538464, "importance_and_relevance": 0.23076923076923078, "materials_and_methods": 0.46153846153846156, "praise": 0.11538461538461539, "presentation_and_reporting": 0, "results_and_discussion": 0.19230769230769232, "suggestion_and_solution": 0.2...	1.576923	iclr2013	openreview	null
SSnY462CYz1Cu	Knowledge Matters: Importance of Prior Information for Optimization	We explore the effect of introducing prior information into the intermediate level of neural networks for a learning task on which all the state-of-the-art machine learning algorithms tested failed to learn. We motivate our work from the hypothesis that humans learn such intermediate concepts from other individuals via...	Knowledge Matters: Importance of Prior Information for Optimization C ¸ a˘ glar G¨ ul¸ cehre gulcehrc@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ erationnelle Universit´ e de Montr´ eal, Montr´ eal, QC, Canada Yoshua Bengio bengioy@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ ...	Çağlar Gülçehre, Yoshua Bengio	Unknown	2,013	{"id": "SSnY462CYz1Cu", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358528400000, "tmdate": 1358528400000, "ddate": null, "number": 36, "content": {"title": "Knowledge Matters: Importance of Prior Information for Optimization", "decision": "conferenceOral-iclr2013-conference...	[Review]: We have uploaded the revision of the paper to arxiv. The revision will be announced by Arxiv soon.	Çağlar Gülçehre	null	null	{"id": "lLgil9MwiZ3Vu", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363246680000, "tmdate": 1363246680000, "ddate": null, "number": 4, "content": {"title": "", "review": "We have uploaded the revision of the paper to arxiv. The revision will be announced by Arxiv soon."}, "f...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 2 }	0	-1.909674	1.909674	0.000774	0.006707	0.000774	0	0	0	0	0	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	0	iclr2013	openreview	null
SSnY462CYz1Cu	Knowledge Matters: Importance of Prior Information for Optimization	We explore the effect of introducing prior information into the intermediate level of neural networks for a learning task on which all the state-of-the-art machine learning algorithms tested failed to learn. We motivate our work from the hypothesis that humans learn such intermediate concepts from other individuals via...	Knowledge Matters: Importance of Prior Information for Optimization C ¸ a˘ glar G¨ ul¸ cehre gulcehrc@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ erationnelle Universit´ e de Montr´ eal, Montr´ eal, QC, Canada Yoshua Bengio bengioy@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ ...	Çağlar Gülçehre, Yoshua Bengio	Unknown	2,013	{"id": "SSnY462CYz1Cu", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358528400000, "tmdate": 1358528400000, "ddate": null, "number": 36, "content": {"title": "Knowledge Matters: Importance of Prior Information for Optimization", "decision": "conferenceOral-iclr2013-conference...	[Review]: The paper by Gulcehre & Bengio entitled 'Knowledge Matters: Importance of Prior Information for Optimization' presents an empirical study which compares a two-tiered MLP architecture against traditional algorithms including SVM, decision trees and boosting. Images used for this task are 64x64 pixel images con...	anonymous reviewer 858d	null	null	{"id": "TiDHTEGclh1ro", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362381600000, "tmdate": 1362381600000, "ddate": null, "number": 1, "content": {"title": "review of Knowledge Matters: Importance of Prior Information for Optimization", "review": "The paper by Gulcehre & Ben...	{ "criticism": 4, "example": 0, "importance_and_relevance": 2, "materials_and_methods": 8, "praise": 3, "presentation_and_reporting": 5, "results_and_discussion": 4, "suggestion_and_solution": 2, "total": 15 }	1.866667	1.803537	0.06313	1.879256	0.161149	0.012589	0.266667	0	0.133333	0.533333	0.2	0.333333	0.266667	0.133333	{ "criticism": 0.26666666666666666, "example": 0, "importance_and_relevance": 0.13333333333333333, "materials_and_methods": 0.5333333333333333, "praise": 0.2, "presentation_and_reporting": 0.3333333333333333, "results_and_discussion": 0.26666666666666666, "suggestion_and_solution": 0.13333333333333333 }	1.866667	iclr2013	openreview	null
SSnY462CYz1Cu	Knowledge Matters: Importance of Prior Information for Optimization	We explore the effect of introducing prior information into the intermediate level of neural networks for a learning task on which all the state-of-the-art machine learning algorithms tested failed to learn. We motivate our work from the hypothesis that humans learn such intermediate concepts from other individuals via...	Knowledge Matters: Importance of Prior Information for Optimization C ¸ a˘ glar G¨ ul¸ cehre gulcehrc@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ erationnelle Universit´ e de Montr´ eal, Montr´ eal, QC, Canada Yoshua Bengio bengioy@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ ...	Çağlar Gülçehre, Yoshua Bengio	Unknown	2,013	{"id": "SSnY462CYz1Cu", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358528400000, "tmdate": 1358528400000, "ddate": null, "number": 36, "content": {"title": "Knowledge Matters: Importance of Prior Information for Optimization", "decision": "conferenceOral-iclr2013-conference...	[Review]: Replies for the reviewers' comments are prepared by the both authors of the paper: Yoshua Bengio and Caglar Gulcehre.	Çağlar Gülçehre	null	null	{"id": "L8RreQWdPS3jz", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363278840000, "tmdate": 1363278840000, "ddate": null, "number": 2, "content": {"title": "", "review": "Replies for the reviewers' comments are prepared by the both authors of the paper: Yoshua Bengio and Cag...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 1 }	0	-2.27267	2.27267	0.000711	0.006163	0.000711	0	0	0	0	0	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	0	iclr2013	openreview	null
SSnY462CYz1Cu	Knowledge Matters: Importance of Prior Information for Optimization	We explore the effect of introducing prior information into the intermediate level of neural networks for a learning task on which all the state-of-the-art machine learning algorithms tested failed to learn. We motivate our work from the hypothesis that humans learn such intermediate concepts from other individuals via...	Knowledge Matters: Importance of Prior Information for Optimization C ¸ a˘ glar G¨ ul¸ cehre gulcehrc@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ erationnelle Universit´ e de Montr´ eal, Montr´ eal, QC, Canada Yoshua Bengio bengioy@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ ...	Çağlar Gülçehre, Yoshua Bengio	Unknown	2,013	{"id": "SSnY462CYz1Cu", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358528400000, "tmdate": 1358528400000, "ddate": null, "number": 36, "content": {"title": "Knowledge Matters: Importance of Prior Information for Optimization", "decision": "conferenceOral-iclr2013-conference...	[Review]: The paper give an example of a task that neural net solves perfectly when intermediate labels are provided but that is not solved at all by several machine learning algorithms including neural net when the intermediate labels are not provided. I consider the result important. Comments: It is surprising th...	anonymous reviewer ed64	null	null	{"id": "D5ft5XCZd1cZw", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361980800000, "tmdate": 1361980800000, "ddate": null, "number": 5, "content": {"title": "review of Knowledge Matters: Importance of Prior Information for Optimization", "review": "The paper give an example o...	{ "criticism": 5, "example": 1, "importance_and_relevance": 0, "materials_and_methods": 5, "praise": 1, "presentation_and_reporting": 6, "results_and_discussion": 2, "suggestion_and_solution": 4, "total": 17 }	1.411765	1.159246	0.252519	1.421626	0.134913	0.009862	0.294118	0.058824	0	0.294118	0.058824	0.352941	0.117647	0.235294	{ "criticism": 0.29411764705882354, "example": 0.058823529411764705, "importance_and_relevance": 0, "materials_and_methods": 0.29411764705882354, "praise": 0.058823529411764705, "presentation_and_reporting": 0.35294117647058826, "results_and_discussion": 0.11764705882352941, "suggestion_and_solution": 0...	1.411765	iclr2013	openreview	null
SSnY462CYz1Cu	Knowledge Matters: Importance of Prior Information for Optimization	We explore the effect of introducing prior information into the intermediate level of neural networks for a learning task on which all the state-of-the-art machine learning algorithms tested failed to learn. We motivate our work from the hypothesis that humans learn such intermediate concepts from other individuals via...	Knowledge Matters: Importance of Prior Information for Optimization C ¸ a˘ glar G¨ ul¸ cehre gulcehrc@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ erationnelle Universit´ e de Montr´ eal, Montr´ eal, QC, Canada Yoshua Bengio bengioy@iro.umontreal.ca D´ epartement d’informatique et de recherche op´ ...	Çağlar Gülçehre, Yoshua Bengio	Unknown	2,013	{"id": "SSnY462CYz1Cu", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358528400000, "tmdate": 1358528400000, "ddate": null, "number": 36, "content": {"title": "Knowledge Matters: Importance of Prior Information for Optimization", "decision": "conferenceOral-iclr2013-conference...	[Review]: In this paper, the authors provide an exposition of curriculum learning and cultural evolution as solutions to the effective local minimum problem. The authors provide a detailed set of simulations that support a curriculum theory of learning, which rely on a supervisory training signal of intermediate task ...	anonymous reviewer dfef	null	null	{"id": "6s7Ys8Q5JbfHZ", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362262800000, "tmdate": 1362262800000, "ddate": null, "number": 6, "content": {"title": "review of Knowledge Matters: Importance of Prior Information for Optimization", "review": "In this paper, the authors ...	{ "criticism": 2, "example": 1, "importance_and_relevance": 4, "materials_and_methods": 12, "praise": 2, "presentation_and_reporting": 6, "results_and_discussion": 6, "suggestion_and_solution": 5, "total": 32 }	1.1875	-4.509957	5.697457	1.20725	0.219048	0.01975	0.0625	0.03125	0.125	0.375	0.0625	0.1875	0.1875	0.15625	{ "criticism": 0.0625, "example": 0.03125, "importance_and_relevance": 0.125, "materials_and_methods": 0.375, "praise": 0.0625, "presentation_and_reporting": 0.1875, "results_and_discussion": 0.1875, "suggestion_and_solution": 0.15625 }	1.1875	iclr2013	openreview	null
PRuOK_LY_WPIq	Matrix Approximation under Local Low-Rank Assumption	Matrix approximation is a common tool in machine learning for building accurate prediction models for recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model w...	Matrix Approximation under Local Low-Rank Assumption Joonseok Leea, Seungyeon Kima, Guy Lebanona, b, Yoram Singerb a College of Computing, Georgia Institute of Technology, Atlanta, GA 30332 b Google Research, Mountain View, CA 94043 {jlee716@, seungyeon.kim@, lebanon@cc.}gatech.edu, singer@google.com Abstract Matrix ap...	Joonseok Lee, Seungyeon Kim, Guy Lebanon, Yoram Singer	Unknown	2,013	{"id": "PRuOK_LY_WPIq", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358317800000, "tmdate": 1358317800000, "ddate": null, "number": 15, "content": {"title": "Matrix Approximation under Local Low-Rank Assumption", "decision": "conferencePoster-iclr2013-workshop", "abstract": ...	[Review]: It has already been mentioned above, but I checked the longer version of the document posted at http://www.cc.gatech.edu/~lebanon/papers/lee_icml_2013.pdf and there really is not enough discussion of the huge previous literature on locally low rank representations, going back at least as far back as http:/...	simon bolivar	null	null	{"id": "JNpPfPeAkDJqK", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363672020000, "tmdate": 1363672020000, "ddate": null, "number": 2, "content": {"title": "", "review": "It has already been mentioned above, but I checked the longer version of the document posted at http://w...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 2, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 3 }	1	-0.578243	1.578243	1.025388	0.230115	0.025388	0.333333	0	0	0.666667	0	0	0	0	{ "criticism": 0.3333333333333333, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.6666666666666666, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	1	iclr2013	openreview	null
PRuOK_LY_WPIq	Matrix Approximation under Local Low-Rank Assumption	Matrix approximation is a common tool in machine learning for building accurate prediction models for recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model w...	Matrix Approximation under Local Low-Rank Assumption Joonseok Leea, Seungyeon Kima, Guy Lebanona, b, Yoram Singerb a College of Computing, Georgia Institute of Technology, Atlanta, GA 30332 b Google Research, Mountain View, CA 94043 {jlee716@, seungyeon.kim@, lebanon@cc.}gatech.edu, singer@google.com Abstract Matrix ap...	Joonseok Lee, Seungyeon Kim, Guy Lebanon, Yoram Singer	Unknown	2,013	{"id": "PRuOK_LY_WPIq", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358317800000, "tmdate": 1358317800000, "ddate": null, "number": 15, "content": {"title": "Matrix Approximation under Local Low-Rank Assumption", "decision": "conferencePoster-iclr2013-workshop", "abstract": ...	[Review]: We appreciate for both of your reviews and questions. - Kernel width: we validated the kernel width experimentally. Specifically, we examined the following kernel types: Gaussian, triangular, and Epanchnikov kernels. We also experimented with the kernel width (0.6, 0.7, 0.8). We found that sufficiently lar...	Joonseok Lee, Seungyeon Kim, Guy Lebanon, Yoram Singer	null	null	{"id": "CkupCgw-sY1o7", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363319940000, "tmdate": 1363319940000, "ddate": null, "number": 1, "content": {"title": "", "review": "We appreciate for both of your reviews and questions.\r\n\r\n- Kernel width: we validated the kernel wid...	{ "criticism": 4, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 11, "praise": 2, "presentation_and_reporting": 4, "results_and_discussion": 6, "suggestion_and_solution": 2, "total": 17 }	1.705882	1.453363	0.252519	1.765131	0.548391	0.059249	0.235294	0	0	0.647059	0.117647	0.235294	0.352941	0.117647	{ "criticism": 0.23529411764705882, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.6470588235294118, "praise": 0.11764705882352941, "presentation_and_reporting": 0.23529411764705882, "results_and_discussion": 0.35294117647058826, "suggestion_and_solution": 0.11764705882352941 }	1.705882	iclr2013	openreview	null
PRuOK_LY_WPIq	Matrix Approximation under Local Low-Rank Assumption	Matrix approximation is a common tool in machine learning for building accurate prediction models for recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model w...	Matrix Approximation under Local Low-Rank Assumption Joonseok Leea, Seungyeon Kima, Guy Lebanona, b, Yoram Singerb a College of Computing, Georgia Institute of Technology, Atlanta, GA 30332 b Google Research, Mountain View, CA 94043 {jlee716@, seungyeon.kim@, lebanon@cc.}gatech.edu, singer@google.com Abstract Matrix ap...	Joonseok Lee, Seungyeon Kim, Guy Lebanon, Yoram Singer	Unknown	2,013	{"id": "PRuOK_LY_WPIq", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358317800000, "tmdate": 1358317800000, "ddate": null, "number": 15, "content": {"title": "Matrix Approximation under Local Low-Rank Assumption", "decision": "conferencePoster-iclr2013-workshop", "abstract": ...	[Review]: Approximation and completion of sparse matrices is a common task. As popularized by the Netflix prize, there are many possible approaches, and combinations of different styles of approach can lead to better predictions than individual methods. In this work, local prediction and low-rank factorization are comb...	anonymous reviewer 76ef	null	null	{"id": "9QsSQSzMpW9Ac", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362191520000, "tmdate": 1362191520000, "ddate": null, "number": 4, "content": {"title": "review of Matrix Approximation under Local Low-Rank Assumption", "review": "Approximation and completion of sparse mat...	{ "criticism": 5, "example": 0, "importance_and_relevance": 2, "materials_and_methods": 8, "praise": 2, "presentation_and_reporting": 5, "results_and_discussion": 7, "suggestion_and_solution": 5, "total": 20 }	1.7	0.926661	0.773339	1.760695	0.59073	0.060695	0.25	0	0.1	0.4	0.1	0.25	0.35	0.25	{ "criticism": 0.25, "example": 0, "importance_and_relevance": 0.1, "materials_and_methods": 0.4, "praise": 0.1, "presentation_and_reporting": 0.25, "results_and_discussion": 0.35, "suggestion_and_solution": 0.25 }	1.7	iclr2013	openreview	null
PRuOK_LY_WPIq	Matrix Approximation under Local Low-Rank Assumption	Matrix approximation is a common tool in machine learning for building accurate prediction models for recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model w...	Matrix Approximation under Local Low-Rank Assumption Joonseok Leea, Seungyeon Kima, Guy Lebanona, b, Yoram Singerb a College of Computing, Georgia Institute of Technology, Atlanta, GA 30332 b Google Research, Mountain View, CA 94043 {jlee716@, seungyeon.kim@, lebanon@cc.}gatech.edu, singer@google.com Abstract Matrix ap...	Joonseok Lee, Seungyeon Kim, Guy Lebanon, Yoram Singer	Unknown	2,013	{"id": "PRuOK_LY_WPIq", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358317800000, "tmdate": 1358317800000, "ddate": null, "number": 15, "content": {"title": "Matrix Approximation under Local Low-Rank Assumption", "decision": "conferencePoster-iclr2013-workshop", "abstract": ...	[Review]: Matrix Approximation under Local Low-Rank Assumption Paper summary This paper deals with low-rank matrix approximation/completion. To reconstruct a matrix element M_{i,j}, the proposed method performs a weighted low rank matrix approximation which considers a similarity metric between matrix coordinates...	anonymous reviewer 4b7c	null	null	{"id": "4eqD-9JEKn4Ea", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362123600000, "tmdate": 1362123600000, "ddate": null, "number": 3, "content": {"title": "review of Matrix Approximation under Local Low-Rank Assumption", "review": "Matrix Approximation under Local Low-Rank ...	{ "criticism": 3, "example": 1, "importance_and_relevance": 1, "materials_and_methods": 17, "praise": 3, "presentation_and_reporting": 9, "results_and_discussion": 10, "suggestion_and_solution": 14, "total": 25 }	2.32	0.04733	2.27267	2.407618	0.870847	0.087618	0.12	0.04	0.04	0.68	0.12	0.36	0.4	0.56	{ "criticism": 0.12, "example": 0.04, "importance_and_relevance": 0.04, "materials_and_methods": 0.68, "praise": 0.12, "presentation_and_reporting": 0.36, "results_and_discussion": 0.4, "suggestion_and_solution": 0.56 }	2.32	iclr2013	openreview	null
OznsOsb6sDFeV	Unsupervised Feature Learning for low-level Local Image Descriptors	Unsupervised feature learning has shown impressive results for a wide range of input modalities, in particular for object classification tasks in computer vision. Using a large amount of unlabeled data, unsupervised feature learning methods are utilized to construct high-level representations that are discriminative en...	Unsupervised Feature Learning for low-level Local Image Descriptors Christian Osendorfer, Justin Bayer, Sebastian Urban, Patrick van der Smagt Technische Universit¨at M¨unchen {osendorf, bayerj, surban, smagt}@in.tum.de Abstract Unsupervised feature learning has shown impressive results for a wide range of in- put moda...	Christian Osendorfer, Justin Bayer, Patrick van der Smagt	Unknown	2,013	{"id": "OznsOsb6sDFeV", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358720100000, "tmdate": 1358720100000, "ddate": null, "number": 17, "content": {"title": "Unsupervised Feature Learning for low-level Local Image Descriptors", "decision": "conferencePoster-iclr2013-workshop...	[Review]: This paper proposes to evaluate feature learning algorithms by using a low-level vision task, namely image patch matching. The authors compare three feature learning algorithms, GRBM. spGRBM and mcRBM against engineered features like SIFT and others. The empirical results unfortunately show that the learned...	anonymous reviewer e954	null	null	{"id": "llHR9RITMyCTz", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362057120000, "tmdate": 1362057120000, "ddate": null, "number": 2, "content": {"title": "review of Unsupervised Feature Learning for low-level Local Image Descriptors", "review": "This paper proposes to eval...	{ "criticism": 12, "example": 1, "importance_and_relevance": 4, "materials_and_methods": 23, "praise": 4, "presentation_and_reporting": 4, "results_and_discussion": 6, "suggestion_and_solution": 9, "total": 37 }	1.702703	-7.387977	9.09068	1.747458	0.452355	0.044756	0.324324	0.027027	0.108108	0.621622	0.108108	0.108108	0.162162	0.243243	{ "criticism": 0.32432432432432434, "example": 0.02702702702702703, "importance_and_relevance": 0.10810810810810811, "materials_and_methods": 0.6216216216216216, "praise": 0.10810810810810811, "presentation_and_reporting": 0.10810810810810811, "results_and_discussion": 0.16216216216216217, "suggestion_a...	1.702703	iclr2013	openreview	null
OznsOsb6sDFeV	Unsupervised Feature Learning for low-level Local Image Descriptors	Unsupervised feature learning has shown impressive results for a wide range of input modalities, in particular for object classification tasks in computer vision. Using a large amount of unlabeled data, unsupervised feature learning methods are utilized to construct high-level representations that are discriminative en...	Unsupervised Feature Learning for low-level Local Image Descriptors Christian Osendorfer, Justin Bayer, Sebastian Urban, Patrick van der Smagt Technische Universit¨at M¨unchen {osendorf, bayerj, surban, smagt}@in.tum.de Abstract Unsupervised feature learning has shown impressive results for a wide range of in- put moda...	Christian Osendorfer, Justin Bayer, Patrick van der Smagt	Unknown	2,013	{"id": "OznsOsb6sDFeV", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358720100000, "tmdate": 1358720100000, "ddate": null, "number": 17, "content": {"title": "Unsupervised Feature Learning for low-level Local Image Descriptors", "decision": "conferencePoster-iclr2013-workshop...	[Review]: his paper proposes a dataset to benchmark the correspodence problem in computer vision. The dataset consists of image patches that have groundtruth matching pairs (using separate algorithms). Extensive experiments show that RBMs perform well compared to hand-crafted features. I like the idea of using i...	anonymous reviewer f716	null	null	{"id": "Hu7OueWCO4ur9", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361947080000, "tmdate": 1361947080000, "ddate": null, "number": 1, "content": {"title": "review of Unsupervised Feature Learning for low-level Local Image Descriptors", "review": "his paper proposes a datase...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 7, "praise": 1, "presentation_and_reporting": 0, "results_and_discussion": 2, "suggestion_and_solution": 1, "total": 7 }	1.571429	1.003261	0.568167	1.581606	0.125383	0.010178	0	0	0	1	0.142857	0	0.285714	0.142857	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 1, "praise": 0.14285714285714285, "presentation_and_reporting": 0, "results_and_discussion": 0.2857142857142857, "suggestion_and_solution": 0.14285714285714285 }	1.571429	iclr2013	openreview	null
OznsOsb6sDFeV	Unsupervised Feature Learning for low-level Local Image Descriptors	Unsupervised feature learning has shown impressive results for a wide range of input modalities, in particular for object classification tasks in computer vision. Using a large amount of unlabeled data, unsupervised feature learning methods are utilized to construct high-level representations that are discriminative en...	Unsupervised Feature Learning for low-level Local Image Descriptors Christian Osendorfer, Justin Bayer, Sebastian Urban, Patrick van der Smagt Technische Universit¨at M¨unchen {osendorf, bayerj, surban, smagt}@in.tum.de Abstract Unsupervised feature learning has shown impressive results for a wide range of in- put moda...	Christian Osendorfer, Justin Bayer, Patrick van der Smagt	Unknown	2,013	{"id": "OznsOsb6sDFeV", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358720100000, "tmdate": 1358720100000, "ddate": null, "number": 17, "content": {"title": "Unsupervised Feature Learning for low-level Local Image Descriptors", "decision": "conferencePoster-iclr2013-workshop...	[Review]: This paper is a survey of unsupervised learning techniques applied to the unsupervised task of descriptor matching. Various methods such as Gaussian RBMs, sparse RBMs, and mcRBMs were applied to image patches and the resulting feature vectors were used in a matching task. These methods were compared to standa...	anonymous reviewer 3338	null	null	{"id": "3wmH3H7ucKwu0", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1361968260000, "tmdate": 1361968260000, "ddate": null, "number": 3, "content": {"title": "review of Unsupervised Feature Learning for low-level Local Image Descriptors", "review": "This paper is a survey of u...	{ "criticism": 4, "example": 1, "importance_and_relevance": 2, "materials_and_methods": 18, "praise": 0, "presentation_and_reporting": 3, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 25 }	1.12	-1.15267	2.27267	1.151751	0.294482	0.031751	0.16	0.04	0.08	0.72	0	0.12	0	0	{ "criticism": 0.16, "example": 0.04, "importance_and_relevance": 0.08, "materials_and_methods": 0.72, "praise": 0, "presentation_and_reporting": 0.12, "results_and_discussion": 0, "suggestion_and_solution": 0 }	1.12	iclr2013	openreview	null
OpvgONa-3WODz	Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines	This paper introduces the Metric-Free Natural Gradient (MFNG) algorithm for training Boltzmann Machines. Similar in spirit to the Hessian-Free method of Martens [8], our algorithm belongs to the family of truncated Newton methods and exploits an efficient matrix-vector product to avoid explicitely storing the natural g...	Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines Guillaume Desjardins, Razvan Pascanu, Aaron Courville and Yoshua Bengio D´epartement d’informatique et de recherche op´erationnelle Universit´e de Montr´eal Abstract This paper introduces the Metric-Free Natural Gradient (MFNG) algorithm for training...	Guillaume Desjardins, Razvan Pascanu, Aaron Courville, Yoshua Bengio	Unknown	2,013	{"id": "OpvgONa-3WODz", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358406900000, "tmdate": 1358406900000, "ddate": null, "number": 65, "content": {"title": "Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines", "decision": "conferencePoster-iclr2013-confer...	[Review]: Thank you to the reviewers for the helpful feedback. The provided references will no doubt come in handy for future work. To all reviewers:In an effort to speedup run time, we have re-implemented a significant portion of the MFNG algorithm. This resulted in large speedups for the diagonal approximation of ...	Guillaume Desjardins, Razvan Pascanu, Aaron Courville, Yoshua Bengio	null	null	{"id": "pC-4pGPkfMnuQ", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363459200000, "tmdate": 1363459200000, "ddate": null, "number": 2, "content": {"title": "", "review": "Thank you to the reviewers for the helpful feedback. The provided references will no doubt come in handy...	{ "criticism": 3, "example": 2, "importance_and_relevance": 1, "materials_and_methods": 19, "praise": 2, "presentation_and_reporting": 6, "results_and_discussion": 8, "suggestion_and_solution": 4, "total": 30 }	1.5	-3.061122	4.561122	1.566432	0.614297	0.066432	0.1	0.066667	0.033333	0.633333	0.066667	0.2	0.266667	0.133333	{ "criticism": 0.1, "example": 0.06666666666666667, "importance_and_relevance": 0.03333333333333333, "materials_and_methods": 0.6333333333333333, "praise": 0.06666666666666667, "presentation_and_reporting": 0.2, "results_and_discussion": 0.26666666666666666, "suggestion_and_solution": 0.1333333333333333...	1.5	iclr2013	openreview	null
OpvgONa-3WODz	Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines	This paper introduces the Metric-Free Natural Gradient (MFNG) algorithm for training Boltzmann Machines. Similar in spirit to the Hessian-Free method of Martens [8], our algorithm belongs to the family of truncated Newton methods and exploits an efficient matrix-vector product to avoid explicitely storing the natural g...	Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines Guillaume Desjardins, Razvan Pascanu, Aaron Courville and Yoshua Bengio D´epartement d’informatique et de recherche op´erationnelle Universit´e de Montr´eal Abstract This paper introduces the Metric-Free Natural Gradient (MFNG) algorithm for training...	Guillaume Desjardins, Razvan Pascanu, Aaron Courville, Yoshua Bengio	Unknown	2,013	{"id": "OpvgONa-3WODz", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358406900000, "tmdate": 1358406900000, "ddate": null, "number": 65, "content": {"title": "Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines", "decision": "conferencePoster-iclr2013-confer...	[Review]: This paper presents a natural gradient algorithm for deep Boltzmann machines. The authors must be commended for their extremely clear and succinct description of the natural gradient method in Section 2. This presentation is particularly useful because, indeed, many of the papers on information geometry are h...	anonymous reviewer 7e2e	null	null	{"id": "o5qvoxIkjTokQ", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362294960000, "tmdate": 1362294960000, "ddate": null, "number": 1, "content": {"title": "review of Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines", "review": "This paper presents a nat...	{ "criticism": 3, "example": 0, "importance_and_relevance": 2, "materials_and_methods": 15, "praise": 4, "presentation_and_reporting": 12, "results_and_discussion": 10, "suggestion_and_solution": 13, "total": 35 }	1.685714	-5.952982	7.638696	1.743111	0.575917	0.057397	0.085714	0	0.057143	0.428571	0.114286	0.342857	0.285714	0.371429	{ "criticism": 0.08571428571428572, "example": 0, "importance_and_relevance": 0.05714285714285714, "materials_and_methods": 0.42857142857142855, "praise": 0.11428571428571428, "presentation_and_reporting": 0.34285714285714286, "results_and_discussion": 0.2857142857142857, "suggestion_and_solution": 0.37...	1.685714	iclr2013	openreview	null
OpvgONa-3WODz	Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines	This paper introduces the Metric-Free Natural Gradient (MFNG) algorithm for training Boltzmann Machines. Similar in spirit to the Hessian-Free method of Martens [8], our algorithm belongs to the family of truncated Newton methods and exploits an efficient matrix-vector product to avoid explicitely storing the natural g...	Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines Guillaume Desjardins, Razvan Pascanu, Aaron Courville and Yoshua Bengio D´epartement d’informatique et de recherche op´erationnelle Universit´e de Montr´eal Abstract This paper introduces the Metric-Free Natural Gradient (MFNG) algorithm for training...	Guillaume Desjardins, Razvan Pascanu, Aaron Courville, Yoshua Bengio	Unknown	2,013	{"id": "OpvgONa-3WODz", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358406900000, "tmdate": 1358406900000, "ddate": null, "number": 65, "content": {"title": "Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines", "decision": "conferencePoster-iclr2013-confer...	[Review]: This paper introduces a new gradient descent algorithm that combines is based on Hessian-free optimization, but replaces the approximate Hessian-vector product by an approximate Fisher information matrix-vector product. It is used to train a DBM, faster than the baseline algorithm in terms of epochs needed, b...	anonymous reviewer 77a7	null	null	{"id": "dt6KtywBaEvBC", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362379800000, "tmdate": 1362379800000, "ddate": null, "number": 3, "content": {"title": "review of Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines", "review": "This paper introduces a n...	{ "criticism": 5, "example": 1, "importance_and_relevance": 1, "materials_and_methods": 8, "praise": 2, "presentation_and_reporting": 7, "results_and_discussion": 4, "suggestion_and_solution": 5, "total": 13 }	2.538462	2.538462	0	2.571649	0.378508	0.033188	0.384615	0.076923	0.076923	0.615385	0.153846	0.538462	0.307692	0.384615	{ "criticism": 0.38461538461538464, "example": 0.07692307692307693, "importance_and_relevance": 0.07692307692307693, "materials_and_methods": 0.6153846153846154, "praise": 0.15384615384615385, "presentation_and_reporting": 0.5384615384615384, "results_and_discussion": 0.3076923076923077, "suggestion_and...	2.538462	iclr2013	openreview	null
OpvgONa-3WODz	Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines	This paper introduces the Metric-Free Natural Gradient (MFNG) algorithm for training Boltzmann Machines. Similar in spirit to the Hessian-Free method of Martens [8], our algorithm belongs to the family of truncated Newton methods and exploits an efficient matrix-vector product to avoid explicitely storing the natural g...	Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines Guillaume Desjardins, Razvan Pascanu, Aaron Courville and Yoshua Bengio D´epartement d’informatique et de recherche op´erationnelle Universit´e de Montr´eal Abstract This paper introduces the Metric-Free Natural Gradient (MFNG) algorithm for training...	Guillaume Desjardins, Razvan Pascanu, Aaron Courville, Yoshua Bengio	Unknown	2,013	{"id": "OpvgONa-3WODz", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358406900000, "tmdate": 1358406900000, "ddate": null, "number": 65, "content": {"title": "Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines", "decision": "conferencePoster-iclr2013-confer...	[Review]: The paper describes a Natural Gradient technique to train Boltzman machines. This is essentially the approach of Amari et al (1992) where the Fisher information matrix is expressed in which the authors estimate the Fisher information matrix L with examples sampled from the model distribution using a MCMC app...	anonymous reviewer 9212	null	null	{"id": "LkyqLtotdQLG4", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362012600000, "tmdate": 1362012600000, "ddate": null, "number": 4, "content": {"title": "review of Metric-Free Natural Gradient for Joint-Training of Boltzmann Machines", "review": "The paper describes a Nat...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 6, "praise": 2, "presentation_and_reporting": 0, "results_and_discussion": 3, "suggestion_and_solution": 0, "total": 9 }	1.333333	1.080814	0.252519	1.353271	0.186196	0.019938	0.111111	0	0	0.666667	0.222222	0	0.333333	0	{ "criticism": 0.1111111111111111, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.6666666666666666, "praise": 0.2222222222222222, "presentation_and_reporting": 0, "results_and_discussion": 0.3333333333333333, "suggestion_and_solution": 0 }	1.333333	iclr2013	openreview	null
OVyHViMbHRm8c	Visual Objects Classification with Sliding Spatial Pyramid Matching	We present a method for visual object classification using only a single feature, transformed color SIFT with a variant of Spatial Pyramid Matching (SPM) that we called Sliding Spatial Pyramid Matching (SSPM), trained with an ensemble of linear regression (provided by LINEAR) to obtained state of the art result on Calt...	arXiv:1212.3767v2 [cs.CV] 18 Dec 2012arXiv: some other text goes here Visual Objects Classiﬁcation with Sliding Spatial Pyramid Matching Hao Wooi Lim 1 and Yong Haur Tay 2 1 haowooilim@1utar.my, 2 tayyh@utar.edu.my Computer Vision & Intelligent Systems (CVIS) group, Faculty of Engineering and Science, University of T...	Hao Wooi Lim, Yong Haur Tay	Unknown	2,013	{"id": "OVyHViMbHRm8c", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358627400000, "tmdate": 1358627400000, "ddate": null, "number": 56, "content": {"title": "Visual Objects Classification with Sliding Spatial Pyramid Matching", "decision": "conferencePoster-iclr2013-workshop...	[Review]: This paper replaces the spatial pyramidal pooling in a spatial pyramid pooling by a sliding-window style pooling. By using this method and color SIFT descriptors, state-of-the-art results are obtained on the Caltech-101 dataset (83.5% accuracy). The contribution in this paper would be rather slight as is,...	anonymous reviewer 9dc6	null	null	{"id": "mqGM7L9xJ-7Oz", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362272400000, "tmdate": 1362272400000, "ddate": null, "number": 2, "content": {"title": "review of Visual Objects Classification with Sliding Spatial Pyramid Matching", "review": "This paper replaces the spa...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 2, "praise": 1, "presentation_and_reporting": 0, "results_and_discussion": 2, "suggestion_and_solution": 0, "total": 3 }	2	0.421757	1.578243	2.016616	0.16406	0.016616	0.333333	0	0	0.666667	0.333333	0	0.666667	0	{ "criticism": 0.3333333333333333, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.6666666666666666, "praise": 0.3333333333333333, "presentation_and_reporting": 0, "results_and_discussion": 0.6666666666666666, "suggestion_and_solution": 0 }	2	iclr2013	openreview	null
OVyHViMbHRm8c	Visual Objects Classification with Sliding Spatial Pyramid Matching	We present a method for visual object classification using only a single feature, transformed color SIFT with a variant of Spatial Pyramid Matching (SPM) that we called Sliding Spatial Pyramid Matching (SSPM), trained with an ensemble of linear regression (provided by LINEAR) to obtained state of the art result on Calt...	arXiv:1212.3767v2 [cs.CV] 18 Dec 2012arXiv: some other text goes here Visual Objects Classiﬁcation with Sliding Spatial Pyramid Matching Hao Wooi Lim 1 and Yong Haur Tay 2 1 haowooilim@1utar.my, 2 tayyh@utar.edu.my Computer Vision & Intelligent Systems (CVIS) group, Faculty of Engineering and Science, University of T...	Hao Wooi Lim, Yong Haur Tay	Unknown	2,013	{"id": "OVyHViMbHRm8c", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358627400000, "tmdate": 1358627400000, "ddate": null, "number": 56, "content": {"title": "Visual Objects Classification with Sliding Spatial Pyramid Matching", "decision": "conferencePoster-iclr2013-workshop...	[Review]: Summary of contributions: The paper presented a method to achieve a state-of-the-art accuracy on the object recognition benchmark Caltech101. The method used two major ingredients: 1. a sliding window of histograms (called sliding spatial pyramid matching) , 2. randomized vocabularies to generate different...	anonymous reviewer 9ba5	null	null	{"id": "-MIjMM8a1GMYx", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362430380000, "tmdate": 1362430380000, "ddate": null, "number": 1, "content": {"title": "review of Visual Objects Classification with Sliding Spatial Pyramid Matching", "review": "Summary of contributions:\r...	{ "criticism": 3, "example": 0, "importance_and_relevance": 3, "materials_and_methods": 5, "praise": 4, "presentation_and_reporting": 3, "results_and_discussion": 2, "suggestion_and_solution": 1, "total": 9 }	2.333333	2.080814	0.252519	2.354945	0.26071	0.021612	0.333333	0	0.333333	0.555556	0.444444	0.333333	0.222222	0.111111	{ "criticism": 0.3333333333333333, "example": 0, "importance_and_relevance": 0.3333333333333333, "materials_and_methods": 0.5555555555555556, "praise": 0.4444444444444444, "presentation_and_reporting": 0.3333333333333333, "results_and_discussion": 0.2222222222222222, "suggestion_and_solution": 0.1111111...	2.333333	iclr2013	openreview	null
OOuGtqpeK-cLI	Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinearities	Recently, we proposed to transform the outputs of each hidden neuron in a multi-layer perceptron network to have zero output and zero slope on average, and use separate shortcut connections to model the linear dependencies instead. We continue the work by firstly introducing a third transformation to normalize the scal...	Pushing Stochastic Gradient towards Second-Order Methods – Backpropagation Learning with Transformations in Nonlinearities Tommi Vatanen, Tapani Raiko, Harri Valpola Department of Information and Computer Science Aalto University School of Science P.O.Box 15400, FI-00076, Aalto, Espoo, Finland first.last@aalto.fi Yann ...	Tommi Vatanen, Tapani Raiko, Harri Valpola, Yann LeCun	Unknown	2,013	{"id": "OOuGtqpeK-cLI", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358347500000, "tmdate": 1358347500000, "ddate": null, "number": 35, "content": {"title": "Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinea...	[Review]: This paper builds on previous work by the same authors that looks at performing dynamic reparameterizations of neural networks to improve training efficiency. The previously published approach is augmented with an additional parameter (gamma) which, although it is argued should help in theory, doesn't seem t...	anonymous reviewer b670	null	null	{"id": "og9azR3sTxoul", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362399720000, "tmdate": 1362399720000, "ddate": null, "number": 1, "content": {"title": "review of Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in...	{ "criticism": 7, "example": 3, "importance_and_relevance": 3, "materials_and_methods": 15, "praise": 3, "presentation_and_reporting": 5, "results_and_discussion": 7, "suggestion_and_solution": 10, "total": 28 }	1.892857	-1.65819	3.551047	1.956411	0.634228	0.063554	0.25	0.107143	0.107143	0.535714	0.107143	0.178571	0.25	0.357143	{ "criticism": 0.25, "example": 0.10714285714285714, "importance_and_relevance": 0.10714285714285714, "materials_and_methods": 0.5357142857142857, "praise": 0.10714285714285714, "presentation_and_reporting": 0.17857142857142858, "results_and_discussion": 0.25, "suggestion_and_solution": 0.35714285714285...	1.892857	iclr2013	openreview	null
OOuGtqpeK-cLI	Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinearities	Recently, we proposed to transform the outputs of each hidden neuron in a multi-layer perceptron network to have zero output and zero slope on average, and use separate shortcut connections to model the linear dependencies instead. We continue the work by firstly introducing a third transformation to normalize the scal...	Pushing Stochastic Gradient towards Second-Order Methods – Backpropagation Learning with Transformations in Nonlinearities Tommi Vatanen, Tapani Raiko, Harri Valpola Department of Information and Computer Science Aalto University School of Science P.O.Box 15400, FI-00076, Aalto, Espoo, Finland first.last@aalto.fi Yann ...	Tommi Vatanen, Tapani Raiko, Harri Valpola, Yann LeCun	Unknown	2,013	{"id": "OOuGtqpeK-cLI", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358347500000, "tmdate": 1358347500000, "ddate": null, "number": 35, "content": {"title": "Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinea...	[Review]: In [10], the authors had previously proposed modifying the network parametrization, in order to ensure zero-mean hidden unit activations across training examples (activity centering) and zero-mean derivatives (slope centering). This was achieved by introducing skip-connections between layers l-1 and l+1 and ...	anonymous reviewer 1567	null	null	{"id": "cAqVvWr0KLv0U", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362183240000, "tmdate": 1362183240000, "ddate": null, "number": 2, "content": {"title": "review of Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in...	{ "criticism": 7, "example": 3, "importance_and_relevance": 3, "materials_and_methods": 12, "praise": 3, "presentation_and_reporting": 6, "results_and_discussion": 6, "suggestion_and_solution": 6, "total": 24 }	1.916667	0.006993	1.909674	1.972295	0.551461	0.055628	0.291667	0.125	0.125	0.5	0.125	0.25	0.25	0.25	{ "criticism": 0.2916666666666667, "example": 0.125, "importance_and_relevance": 0.125, "materials_and_methods": 0.5, "praise": 0.125, "presentation_and_reporting": 0.25, "results_and_discussion": 0.25, "suggestion_and_solution": 0.25 }	1.916667	iclr2013	openreview	null
OOuGtqpeK-cLI	Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinearities	Recently, we proposed to transform the outputs of each hidden neuron in a multi-layer perceptron network to have zero output and zero slope on average, and use separate shortcut connections to model the linear dependencies instead. We continue the work by firstly introducing a third transformation to normalize the scal...	Pushing Stochastic Gradient towards Second-Order Methods – Backpropagation Learning with Transformations in Nonlinearities Tommi Vatanen, Tapani Raiko, Harri Valpola Department of Information and Computer Science Aalto University School of Science P.O.Box 15400, FI-00076, Aalto, Espoo, Finland first.last@aalto.fi Yann ...	Tommi Vatanen, Tapani Raiko, Harri Valpola, Yann LeCun	Unknown	2,013	{"id": "OOuGtqpeK-cLI", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358347500000, "tmdate": 1358347500000, "ddate": null, "number": 35, "content": {"title": "Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinea...	[Review]: * A brief summary of the paper's contributions, in the context of prior work. This paper extends the authors' previous work on making sure that the hidden units in a neural net have zero output and slope on average, by also using direct connections that model explicitly the linear dependencies. The extensi...	anonymous reviewer c3d4	null	null	{"id": "Id_EI3kn5mX4i", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362387060000, "tmdate": 1362387060000, "ddate": null, "number": 3, "content": {"title": "review of Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in...	{ "criticism": 2, "example": 0, "importance_and_relevance": 4, "materials_and_methods": 11, "praise": 4, "presentation_and_reporting": 2, "results_and_discussion": 7, "suggestion_and_solution": 2, "total": 16 }	2	1.857958	0.142042	2.035511	0.371634	0.035511	0.125	0	0.25	0.6875	0.25	0.125	0.4375	0.125	{ "criticism": 0.125, "example": 0, "importance_and_relevance": 0.25, "materials_and_methods": 0.6875, "praise": 0.25, "presentation_and_reporting": 0.125, "results_and_discussion": 0.4375, "suggestion_and_solution": 0.125 }	2	iclr2013	openreview	null
OOuGtqpeK-cLI	Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinearities	Recently, we proposed to transform the outputs of each hidden neuron in a multi-layer perceptron network to have zero output and zero slope on average, and use separate shortcut connections to model the linear dependencies instead. We continue the work by firstly introducing a third transformation to normalize the scal...	Pushing Stochastic Gradient towards Second-Order Methods – Backpropagation Learning with Transformations in Nonlinearities Tommi Vatanen, Tapani Raiko, Harri Valpola Department of Information and Computer Science Aalto University School of Science P.O.Box 15400, FI-00076, Aalto, Espoo, Finland first.last@aalto.fi Yann ...	Tommi Vatanen, Tapani Raiko, Harri Valpola, Yann LeCun	Unknown	2,013	{"id": "OOuGtqpeK-cLI", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358347500000, "tmdate": 1358347500000, "ddate": null, "number": 35, "content": {"title": "Pushing Stochastic Gradient towards Second-Order Methods -- Backpropagation Learning with Transformations in Nonlinea...	[Review]: First of all we would like to thank you for your informed, thorough and kind comments. We realize that there is major overlap with our previous paper [10]. We hope that these two papers could be combined in a journal paper later on. It was mentioned that we use some text verbatim from [10]. There is some basi...	Tommi Vatanen, Tapani Raiko, Harri Valpola, Yann LeCun	null	null	{"id": "8PUQYHnMEx8CL", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363039740000, "tmdate": 1363039740000, "ddate": null, "number": 4, "content": {"title": "", "review": "First of all we would like to thank you for your informed, thorough and kind comments. We realize that t...	{ "criticism": 4, "example": 2, "importance_and_relevance": 2, "materials_and_methods": 10, "praise": 2, "presentation_and_reporting": 14, "results_and_discussion": 10, "suggestion_and_solution": 12, "total": 45 }	1.244444	-14.916764	16.161208	1.308916	0.615857	0.064472	0.088889	0.044444	0.044444	0.222222	0.044444	0.311111	0.222222	0.266667	{ "criticism": 0.08888888888888889, "example": 0.044444444444444446, "importance_and_relevance": 0.044444444444444446, "materials_and_methods": 0.2222222222222222, "praise": 0.044444444444444446, "presentation_and_reporting": 0.3111111111111111, "results_and_discussion": 0.2222222222222222, "suggestion_...	1.244444	iclr2013	openreview	null
N_c1XDpyus_yP	A Nested HDP for Hierarchical Topic Models	We develop a nested hierarchical Dirichlet process (nHDP) for hierarchical topic modeling. The nHDP is a generalization of the nested Chinese restaurant process (nCRP) that allows each word to follow its own path to a topic node according to a document-specific distribution on a shared tree. This alleviates the rigid, ...	A Nested HDP for Hierarchical Topic Models John Paisley Department of EECS UC Berkeley Chong Wang Dept. of Machine Learning Carnegie Mellon University David Blei Dept. of Computer Science Princeton University Michael I. Jordan Department of EECS UC Berkeley Abstract We develop a nested hierarchical Dirichlet process (n...	John Paisley, Chong Wang, David Blei, Michael I. Jordan	Unknown	2,013	{"id": "N_c1XDpyus_yP", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358404200000, "tmdate": 1358404200000, "ddate": null, "number": 22, "content": {"title": "A Nested HDP for Hierarchical Topic Models", "decision": "conferenceOral-iclr2013-workshop", "abstract": "We develop ...	[Review]: The paper introduces a natural extension to the nested Chinese Restaurant process, where the main limitation was that a single path for the tree (from the root to a leaf) is chosen for each individual document. In this work, a document specific tree is drawn (with associated switching probabilities) which is ...	anonymous reviewer 7555	null	null	{"id": "s3Zn3ZANM4Twv", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362170460000, "tmdate": 1362170460000, "ddate": null, "number": 2, "content": {"title": "review of A Nested HDP for Hierarchical Topic Models", "review": "The paper introduces a natural extension to the nest...	{ "criticism": 2, "example": 1, "importance_and_relevance": 2, "materials_and_methods": 3, "praise": 2, "presentation_and_reporting": 6, "results_and_discussion": 2, "suggestion_and_solution": 4, "total": 10 }	2.2	2.057958	0.142042	2.264767	0.663533	0.064767	0.2	0.1	0.2	0.3	0.2	0.6	0.2	0.4	{ "criticism": 0.2, "example": 0.1, "importance_and_relevance": 0.2, "materials_and_methods": 0.3, "praise": 0.2, "presentation_and_reporting": 0.6, "results_and_discussion": 0.2, "suggestion_and_solution": 0.4 }	2.2	iclr2013	openreview	null
N_c1XDpyus_yP	A Nested HDP for Hierarchical Topic Models	We develop a nested hierarchical Dirichlet process (nHDP) for hierarchical topic modeling. The nHDP is a generalization of the nested Chinese restaurant process (nCRP) that allows each word to follow its own path to a topic node according to a document-specific distribution on a shared tree. This alleviates the rigid, ...	A Nested HDP for Hierarchical Topic Models John Paisley Department of EECS UC Berkeley Chong Wang Dept. of Machine Learning Carnegie Mellon University David Blei Dept. of Computer Science Princeton University Michael I. Jordan Department of EECS UC Berkeley Abstract We develop a nested hierarchical Dirichlet process (n...	John Paisley, Chong Wang, David Blei, Michael I. Jordan	Unknown	2,013	{"id": "N_c1XDpyus_yP", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358404200000, "tmdate": 1358404200000, "ddate": null, "number": 22, "content": {"title": "A Nested HDP for Hierarchical Topic Models", "decision": "conferenceOral-iclr2013-workshop", "abstract": "We develop ...	[Review]: no additional comments.	anonymous reviewer 95fc	null	null	{"id": "cBZ06aJhuH6Nw", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362389580000, "tmdate": 1362389580000, "ddate": null, "number": 3, "content": {"title": "", "review": "no additional comments."}, "forum": "N_c1XDpyus_yP", "referent": null, "invitation": "ICLR.cc/2013/-/sub...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 1 }	1	-1.27267	2.27267	1.001445	0.022525	0.001445	1	0	0	0	0	0	0	0	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	1	iclr2013	openreview	null
N_c1XDpyus_yP	A Nested HDP for Hierarchical Topic Models	We develop a nested hierarchical Dirichlet process (nHDP) for hierarchical topic modeling. The nHDP is a generalization of the nested Chinese restaurant process (nCRP) that allows each word to follow its own path to a topic node according to a document-specific distribution on a shared tree. This alleviates the rigid, ...	A Nested HDP for Hierarchical Topic Models John Paisley Department of EECS UC Berkeley Chong Wang Dept. of Machine Learning Carnegie Mellon University David Blei Dept. of Computer Science Princeton University Michael I. Jordan Department of EECS UC Berkeley Abstract We develop a nested hierarchical Dirichlet process (n...	John Paisley, Chong Wang, David Blei, Michael I. Jordan	Unknown	2,013	{"id": "N_c1XDpyus_yP", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358404200000, "tmdate": 1358404200000, "ddate": null, "number": 22, "content": {"title": "A Nested HDP for Hierarchical Topic Models", "decision": "conferenceOral-iclr2013-workshop", "abstract": "We develop ...	[Review]: no additional comments.	anonymous reviewer 95fc	null	null	{"id": "WZaI2aHNOvDz7", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362389640000, "tmdate": 1362389640000, "ddate": null, "number": 1, "content": {"title": "", "review": "no additional comments."}, "forum": "N_c1XDpyus_yP", "referent": null, "invitation": "ICLR.cc/2013/-/sub...	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 1 }	1	-1.27267	2.27267	1.001445	0.022525	0.001445	1	0	0	0	0	0	0	0	{ "criticism": 1, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	1	iclr2013	openreview	null
N_c1XDpyus_yP	A Nested HDP for Hierarchical Topic Models	We develop a nested hierarchical Dirichlet process (nHDP) for hierarchical topic modeling. The nHDP is a generalization of the nested Chinese restaurant process (nCRP) that allows each word to follow its own path to a topic node according to a document-specific distribution on a shared tree. This alleviates the rigid, ...	A Nested HDP for Hierarchical Topic Models John Paisley Department of EECS UC Berkeley Chong Wang Dept. of Machine Learning Carnegie Mellon University David Blei Dept. of Computer Science Princeton University Michael I. Jordan Department of EECS UC Berkeley Abstract We develop a nested hierarchical Dirichlet process (n...	John Paisley, Chong Wang, David Blei, Michael I. Jordan	Unknown	2,013	{"id": "N_c1XDpyus_yP", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358404200000, "tmdate": 1358404200000, "ddate": null, "number": 22, "content": {"title": "A Nested HDP for Hierarchical Topic Models", "decision": "conferenceOral-iclr2013-workshop", "abstract": "We develop ...	[Review]: This paper presents a novel variant of the NCRP process that overcomes the latter's main limitation, namely, that a document necessarily has to use topics from a specific path in the tree. This is accomplished by combining ideas from HDP with the NCRP process, where the entire nCRP tree is replicated for each...	anonymous reviewer 95fc	null	null	{"id": "BAmMaGEF72a0w", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362389460000, "tmdate": 1362389460000, "ddate": null, "number": 4, "content": {"title": "review of A Nested HDP for Hierarchical Topic Models", "review": "This paper presents a novel variant of the NCRP proc...	{ "criticism": 2, "example": 0, "importance_and_relevance": 3, "materials_and_methods": 6, "praise": 2, "presentation_and_reporting": 2, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 9 }	1.666667	1.414148	0.252519	1.712917	0.450993	0.046251	0.222222	0	0.333333	0.666667	0.222222	0.222222	0	0	{ "criticism": 0.2222222222222222, "example": 0, "importance_and_relevance": 0.3333333333333333, "materials_and_methods": 0.6666666666666666, "praise": 0.2222222222222222, "presentation_and_reporting": 0.2222222222222222, "results_and_discussion": 0, "suggestion_and_solution": 0 }	1.666667	iclr2013	openreview	null
MQm0HKx20L7iN	Kernelized Locality-Sensitive Hashing for Semi-Supervised Agglomerative Clustering	Large scale agglomerative clustering is hindered by computational burdens. We propose a novel scheme where exact inter-instance distance calculation is replaced by the Hamming distance between Kernelized Locality-Sensitive Hashing (KLSH) hashed values. This results in a method that drastically decreases computation tim...	arXiv:1301.3575v1 [cs.LG] 16 Jan 2013 Kernelized Locality-Sensitive Hashing for Semi-Supervised Agglomerative Clustering Boyi Xie Department of Computer Science Columbia University New York, NY 10027 xie@cs.columbia.edu Shuheng Zheng Department of Industrial Engineering & Operations Research Columbia University New Y...	Boyi Xie, Shuheng Zheng	Unknown	2,013	{"id": "MQm0HKx20L7iN", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358448300000, "tmdate": 1358448300000, "ddate": null, "number": 26, "content": {"decision": "reject", "title": "Kernelized Locality-Sensitive Hashing for Semi-Supervised Agglomerative Clustering", "abstract"...	[Review]: This paper proposes to use kernelized locality-sensitive hashing (KLSH), based on a similarity metric learned from labeled data, to accelerate agglomerative (hierarchical) clustering. Agglomerative clustering requires, at each iteration, to find the pair of closest clusters. The idea behind this paper is that...	anonymous reviewer c8d7	null	null	{"id": "vpc3vyRo-2AFM", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362080280000, "tmdate": 1362080280000, "ddate": null, "number": 1, "content": {"title": "review of Kernelized Locality-Sensitive Hashing for Semi-Supervised Agglomerative Clustering", "review": "This paper p...	{ "criticism": 11, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 17, "praise": 0, "presentation_and_reporting": 6, "results_and_discussion": 3, "suggestion_and_solution": 2, "total": 26 }	1.5	-1.167231	2.667231	1.573031	0.659713	0.073031	0.423077	0	0	0.653846	0	0.230769	0.115385	0.076923	{ "criticism": 0.4230769230769231, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.6538461538461539, "praise": 0, "presentation_and_reporting": 0.23076923076923078, "results_and_discussion": 0.11538461538461539, "suggestion_and_solution": 0.07692307692307693 }	1.5	iclr2013	openreview	null
MQm0HKx20L7iN	Kernelized Locality-Sensitive Hashing for Semi-Supervised Agglomerative Clustering	Large scale agglomerative clustering is hindered by computational burdens. We propose a novel scheme where exact inter-instance distance calculation is replaced by the Hamming distance between Kernelized Locality-Sensitive Hashing (KLSH) hashed values. This results in a method that drastically decreases computation tim...	arXiv:1301.3575v1 [cs.LG] 16 Jan 2013 Kernelized Locality-Sensitive Hashing for Semi-Supervised Agglomerative Clustering Boyi Xie Department of Computer Science Columbia University New York, NY 10027 xie@cs.columbia.edu Shuheng Zheng Department of Industrial Engineering & Operations Research Columbia University New Y...	Boyi Xie, Shuheng Zheng	Unknown	2,013	{"id": "MQm0HKx20L7iN", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358448300000, "tmdate": 1358448300000, "ddate": null, "number": 26, "content": {"decision": "reject", "title": "Kernelized Locality-Sensitive Hashing for Semi-Supervised Agglomerative Clustering", "abstract"...	[Review]: This workshop submission proposes a method for clustering data which applies a semi-supervised distance metric to the data prior to applying kernelized locality-sensitive hashing for agglom erative clustering. The intuition is that distance learning on a subset of data pairs will improve overall performance,...	anonymous reviewer cce9	null	null	{"id": "Z9bz9yXn_F9nA", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362172860000, "tmdate": 1362172860000, "ddate": null, "number": 2, "content": {"title": "review of Kernelized Locality-Sensitive Hashing for Semi-Supervised Agglomerative Clustering", "review": "This worksho...	{ "criticism": 4, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 7, "praise": 0, "presentation_and_reporting": 1, "results_and_discussion": 3, "suggestion_and_solution": 1, "total": 9 }	1.888889	1.63637	0.252519	1.919705	0.313846	0.030817	0.444444	0	0.111111	0.777778	0	0.111111	0.333333	0.111111	{ "criticism": 0.4444444444444444, "example": 0, "importance_and_relevance": 0.1111111111111111, "materials_and_methods": 0.7777777777777778, "praise": 0, "presentation_and_reporting": 0.1111111111111111, "results_and_discussion": 0.3333333333333333, "suggestion_and_solution": 0.1111111111111111 }	1.888889	iclr2013	openreview	null
KKZ-FeUj-9kjY	Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint	Deep Belief Networks (DBN) have been successfully applied on popular machine learning tasks. Specifically, when applied on hand-written digit recognition, DBNs have achieved approximate accuracy rates of 98.8%. In an effort to optimize the data representation achieved by the DBN and maximize their descriptive power, re...	Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint Xanadu C. Halkias DYNI, LSIS, Universit`e du Sud, Avenue de l’Universit´e - BP20132, 83957 LA GARDE CEDEX - FRANCE xanadu.halkias@univ-tln.fr S´ebastien Paris DYNI, LSIS CNRS UMR 7296, Aix-Marseille University Domaine universitaire de Saint J´erˆom...	Xanadu Halkias, Sébastien PARIS, Herve Glotin	Unknown	2,013	{"id": "KKZ-FeUj-9kjY", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358700300000, "tmdate": 1358700300000, "ddate": null, "number": 50, "content": {"decision": "reject", "title": "Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint", "abstract": "Deep Bel...	[Review]: The paper proposes a mixed norm penalty for regularizing RBMs and DBNs. The work extends previous work on sparse RBMs and DBNs and extends the work of Luo et al. (2011) on sparse group RBMs (and DBMs) to deep belief nets. The method is tested on several datasets and no significant improvement is reported comp...	anonymous reviewer 0136	null	null	{"id": "ttT0L-IGxpbuw", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362153000000, "tmdate": 1362153000000, "ddate": null, "number": 1, "content": {"title": "review of Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint", "review": "The paper proposes a mi...	{ "criticism": 6, "example": 3, "importance_and_relevance": 0, "materials_and_methods": 11, "praise": 0, "presentation_and_reporting": 9, "results_and_discussion": 0, "suggestion_and_solution": 4, "total": 22 }	1.5	0.221623	1.278377	1.53757	0.368002	0.03757	0.272727	0.136364	0	0.5	0	0.409091	0	0.181818	{ "criticism": 0.2727272727272727, "example": 0.13636363636363635, "importance_and_relevance": 0, "materials_and_methods": 0.5, "praise": 0, "presentation_and_reporting": 0.4090909090909091, "results_and_discussion": 0, "suggestion_and_solution": 0.18181818181818182 }	1.5	iclr2013	openreview	null
KKZ-FeUj-9kjY	Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint	Deep Belief Networks (DBN) have been successfully applied on popular machine learning tasks. Specifically, when applied on hand-written digit recognition, DBNs have achieved approximate accuracy rates of 98.8%. In an effort to optimize the data representation achieved by the DBN and maximize their descriptive power, re...	Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint Xanadu C. Halkias DYNI, LSIS, Universit`e du Sud, Avenue de l’Universit´e - BP20132, 83957 LA GARDE CEDEX - FRANCE xanadu.halkias@univ-tln.fr S´ebastien Paris DYNI, LSIS CNRS UMR 7296, Aix-Marseille University Domaine universitaire de Saint J´erˆom...	Xanadu Halkias, Sébastien PARIS, Herve Glotin	Unknown	2,013	{"id": "KKZ-FeUj-9kjY", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358700300000, "tmdate": 1358700300000, "ddate": null, "number": 50, "content": {"decision": "reject", "title": "Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint", "abstract": "Deep Bel...	[Review]: In this paper the authors propose a method to make the hidden units of RBM group sparse. The key idea is to add a penalty term to the negative log-likelihood loss penalizing the L2/L1 norm over the activations of the RBM. The authors demonstrate their method on three digit classification tasks. These experime...	anonymous reviewer 61fc	null	null	{"id": "ijgMjq-uMOiYw", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362144480000, "tmdate": 1362144480000, "ddate": null, "number": 3, "content": {"title": "review of Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint", "review": "In this paper the autho...	{ "criticism": 13, "example": 1, "importance_and_relevance": 2, "materials_and_methods": 24, "praise": 2, "presentation_and_reporting": 8, "results_and_discussion": 12, "suggestion_and_solution": 8, "total": 43 }	1.627907	-12.57628	14.204187	1.704609	0.713844	0.076702	0.302326	0.023256	0.046512	0.55814	0.046512	0.186047	0.27907	0.186047	{ "criticism": 0.3023255813953488, "example": 0.023255813953488372, "importance_and_relevance": 0.046511627906976744, "materials_and_methods": 0.5581395348837209, "praise": 0.046511627906976744, "presentation_and_reporting": 0.18604651162790697, "results_and_discussion": 0.27906976744186046, "suggestion...	1.627907	iclr2013	openreview	null
KKZ-FeUj-9kjY	Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint	Deep Belief Networks (DBN) have been successfully applied on popular machine learning tasks. Specifically, when applied on hand-written digit recognition, DBNs have achieved approximate accuracy rates of 98.8%. In an effort to optimize the data representation achieved by the DBN and maximize their descriptive power, re...	Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint Xanadu C. Halkias DYNI, LSIS, Universit`e du Sud, Avenue de l’Universit´e - BP20132, 83957 LA GARDE CEDEX - FRANCE xanadu.halkias@univ-tln.fr S´ebastien Paris DYNI, LSIS CNRS UMR 7296, Aix-Marseille University Domaine universitaire de Saint J´erˆom...	Xanadu Halkias, Sébastien PARIS, Herve Glotin	Unknown	2,013	{"id": "KKZ-FeUj-9kjY", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358700300000, "tmdate": 1358700300000, "ddate": null, "number": 50, "content": {"decision": "reject", "title": "Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint", "abstract": "Deep Bel...	[Review]: Since the last version of the paper (v2) is incomplete my following comments are mainly based on the first version. This paper proposes using $l_{1,2}$ regularization (for both non-overlapping and overlapping groups) upon the activation possibilities of hidden units in RBMs. Then DBNs pretrained by the res...	anonymous reviewer e6d4	null	null	{"id": "dWSK4E1RkeWRi", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362193620000, "tmdate": 1362193620000, "ddate": null, "number": 2, "content": {"title": "review of Sparse Penalty in Deep Belief Networks: Using the Mixed Norm Constraint", "review": "Since the last version ...	{ "criticism": 4, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 3, "praise": 0, "presentation_and_reporting": 3, "results_and_discussion": 1, "suggestion_and_solution": 1, "total": 7 }	1.714286	1.146118	0.568167	1.734556	0.214318	0.020271	0.571429	0	0	0.428571	0	0.428571	0.142857	0.142857	{ "criticism": 0.5714285714285714, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0.42857142857142855, "praise": 0, "presentation_and_reporting": 0.42857142857142855, "results_and_discussion": 0.14285714285714285, "suggestion_and_solution": 0.14285714285714285 }	1.714286	iclr2013	openreview	null
KHMdKiX2lbguE	Boltzmann Machines and Denoising Autoencoders for Image Denoising	Image denoising based on a probabilistic model of local image patches has been employed by various researchers, and recently a deep (denoising) autoencoder has been proposed by Burger et al. [2012] and Xie et al. [2012] as a good model for this. In this paper, we propose that another popular family of models in the fie...	arXiv:1301.3468v6 [stat.ML] 4 Mar 2013 Boltzmann Machines and Denoising Autoencoders for Image Denoising KyungHyun Cho Aalto University School of Science Department of Information and Computer Science Espoo, Finland kyunghyun.cho@aalto.fi Abstract Image denoising based on a probabilistic model of local imag e patches...	KyungHyun Cho	Unknown	2,013	{"id": "KHMdKiX2lbguE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358345700000, "tmdate": 1358345700000, "ddate": null, "number": 66, "content": {"title": "Boltzmann Machines and Denoising Autoencoders for Image Denoising", "decision": "conferencePoster-iclr2013-workshop",...	[Review]: The paper conducts an empirical performance comparison, on the task of image denoising, where the denoising of large images is based on combining densoing of small patches. In this context, the study compares usign, as small patch denoisers, deep denoising autoencoders (DAE) versus deep Boltzmann machines wit...	anonymous reviewer 9120	null	null	{"id": "tO_8tX3y-7SXz", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362361020000, "tmdate": 1362361020000, "ddate": null, "number": 2, "content": {"title": "review of Boltzmann Machines and Denoising Autoencoders for Image Denoising", "review": "The paper conducts an empiric...	{ "criticism": 5, "example": 1, "importance_and_relevance": 2, "materials_and_methods": 9, "praise": 3, "presentation_and_reporting": 2, "results_and_discussion": 4, "suggestion_and_solution": 3, "total": 17 }	1.705882	1.453363	0.252519	1.738301	0.336364	0.032419	0.294118	0.058824	0.117647	0.529412	0.176471	0.117647	0.235294	0.176471	{ "criticism": 0.29411764705882354, "example": 0.058823529411764705, "importance_and_relevance": 0.11764705882352941, "materials_and_methods": 0.5294117647058824, "praise": 0.17647058823529413, "presentation_and_reporting": 0.11764705882352941, "results_and_discussion": 0.23529411764705882, "suggestion_...	1.705882	iclr2013	openreview	null
KHMdKiX2lbguE	Boltzmann Machines and Denoising Autoencoders for Image Denoising	Image denoising based on a probabilistic model of local image patches has been employed by various researchers, and recently a deep (denoising) autoencoder has been proposed by Burger et al. [2012] and Xie et al. [2012] as a good model for this. In this paper, we propose that another popular family of models in the fie...	arXiv:1301.3468v6 [stat.ML] 4 Mar 2013 Boltzmann Machines and Denoising Autoencoders for Image Denoising KyungHyun Cho Aalto University School of Science Department of Information and Computer Science Espoo, Finland kyunghyun.cho@aalto.fi Abstract Image denoising based on a probabilistic model of local imag e patches...	KyungHyun Cho	Unknown	2,013	{"id": "KHMdKiX2lbguE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358345700000, "tmdate": 1358345700000, "ddate": null, "number": 66, "content": {"title": "Boltzmann Machines and Denoising Autoencoders for Image Denoising", "decision": "conferencePoster-iclr2013-workshop",...	[Review]: Dear reviewers (bf00) and (9120), First of all, thank you for your thorough reviews. Please, find my response to your comments below. A revision of the paper that includes the fixes made accordingly will be available at the arXiv.org tomorrow (Tue, 5 Mar 2013 01:00:00 GMT). To both reviewers (bf0...	Kyunghyun Cho	null	null	{"id": "ppSEYjkaMGYj5", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362411780000, "tmdate": 1362411780000, "ddate": null, "number": 1, "content": {"title": "", "review": "Dear reviewers (bf00) and (9120),\r\n\r\nFirst of all, thank you for your thorough reviews. \r\n\r\nPlea...	{ "criticism": 4, "example": 4, "importance_and_relevance": 1, "materials_and_methods": 12, "praise": 0, "presentation_and_reporting": 11, "results_and_discussion": 4, "suggestion_and_solution": 8, "total": 28 }	1.571429	-1.979618	3.551047	1.604497	0.348849	0.033069	0.142857	0.142857	0.035714	0.428571	0	0.392857	0.142857	0.285714	{ "criticism": 0.14285714285714285, "example": 0.14285714285714285, "importance_and_relevance": 0.03571428571428571, "materials_and_methods": 0.42857142857142855, "praise": 0, "presentation_and_reporting": 0.39285714285714285, "results_and_discussion": 0.14285714285714285, "suggestion_and_solution": 0.2...	1.571429	iclr2013	openreview	null
KHMdKiX2lbguE	Boltzmann Machines and Denoising Autoencoders for Image Denoising	Image denoising based on a probabilistic model of local image patches has been employed by various researchers, and recently a deep (denoising) autoencoder has been proposed by Burger et al. [2012] and Xie et al. [2012] as a good model for this. In this paper, we propose that another popular family of models in the fie...	arXiv:1301.3468v6 [stat.ML] 4 Mar 2013 Boltzmann Machines and Denoising Autoencoders for Image Denoising KyungHyun Cho Aalto University School of Science Department of Information and Computer Science Espoo, Finland kyunghyun.cho@aalto.fi Abstract Image denoising based on a probabilistic model of local imag e patches...	KyungHyun Cho	Unknown	2,013	{"id": "KHMdKiX2lbguE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358345700000, "tmdate": 1358345700000, "ddate": null, "number": 66, "content": {"title": "Boltzmann Machines and Denoising Autoencoders for Image Denoising", "decision": "conferencePoster-iclr2013-workshop",...	[Review]: Dear reviewer (d5d4), Thank you for your thorough review and comments. - 'the paper fails to compare against robust Boltzmann machines (Tang et al., CVPR 2012)' Thanks for pointing it out, and I agree that the RoBM be tried as well. It will be possible to use the already tr...	Kyunghyun Cho	null	null	{"id": "VC6Ay131A-y1w", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362494700000, "tmdate": 1362494700000, "ddate": null, "number": 3, "content": {"title": "", "review": "Dear reviewer (d5d4),\r\n\r\nThank you for your thorough review and comments.\r\n \r\n - 'the paper fail...	{ "criticism": 3, "example": 1, "importance_and_relevance": 1, "materials_and_methods": 13, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 1, "suggestion_and_solution": 3, "total": 17 }	1.294118	1.041599	0.252519	1.317695	0.249709	0.023577	0.176471	0.058824	0.058824	0.764706	0	0	0.058824	0.176471	{ "criticism": 0.17647058823529413, "example": 0.058823529411764705, "importance_and_relevance": 0.058823529411764705, "materials_and_methods": 0.7647058823529411, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0.058823529411764705, "suggestion_and_solution": 0.17647058823529413...	1.294118	iclr2013	openreview	null
KHMdKiX2lbguE	Boltzmann Machines and Denoising Autoencoders for Image Denoising	Image denoising based on a probabilistic model of local image patches has been employed by various researchers, and recently a deep (denoising) autoencoder has been proposed by Burger et al. [2012] and Xie et al. [2012] as a good model for this. In this paper, we propose that another popular family of models in the fie...	arXiv:1301.3468v6 [stat.ML] 4 Mar 2013 Boltzmann Machines and Denoising Autoencoders for Image Denoising KyungHyun Cho Aalto University School of Science Department of Information and Computer Science Espoo, Finland kyunghyun.cho@aalto.fi Abstract Image denoising based on a probabilistic model of local imag e patches...	KyungHyun Cho	Unknown	2,013	{"id": "KHMdKiX2lbguE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358345700000, "tmdate": 1358345700000, "ddate": null, "number": 66, "content": {"title": "Boltzmann Machines and Denoising Autoencoders for Image Denoising", "decision": "conferencePoster-iclr2013-workshop",...	[Review]: This paper is an empirical comparison of the different models (Boltzmann Machines and Denoising Autoencoders) on the task of image denoising. Based on the experiments the authors claimed the increasing model depth improves the denoising performances when the level of noise is high. PROS + Exploring DBMs f...	anonymous reviewer bf00	null	null	{"id": "PLgu8d4J3rRz9", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362189720000, "tmdate": 1362189720000, "ddate": null, "number": 5, "content": {"title": "review of Boltzmann Machines and Denoising Autoencoders for Image Denoising", "review": "This paper is an empirical co...	{ "criticism": 2, "example": 5, "importance_and_relevance": 1, "materials_and_methods": 7, "praise": 1, "presentation_and_reporting": 5, "results_and_discussion": 1, "suggestion_and_solution": 6, "total": 13 }	2.153846	2.153846	0	2.169495	0.234136	0.015649	0.153846	0.384615	0.076923	0.538462	0.076923	0.384615	0.076923	0.461538	{ "criticism": 0.15384615384615385, "example": 0.38461538461538464, "importance_and_relevance": 0.07692307692307693, "materials_and_methods": 0.5384615384615384, "praise": 0.07692307692307693, "presentation_and_reporting": 0.38461538461538464, "results_and_discussion": 0.07692307692307693, "suggestion_a...	2.153846	iclr2013	openreview	null
KHMdKiX2lbguE	Boltzmann Machines and Denoising Autoencoders for Image Denoising	Image denoising based on a probabilistic model of local image patches has been employed by various researchers, and recently a deep (denoising) autoencoder has been proposed by Burger et al. [2012] and Xie et al. [2012] as a good model for this. In this paper, we propose that another popular family of models in the fie...	arXiv:1301.3468v6 [stat.ML] 4 Mar 2013 Boltzmann Machines and Denoising Autoencoders for Image Denoising KyungHyun Cho Aalto University School of Science Department of Information and Computer Science Espoo, Finland kyunghyun.cho@aalto.fi Abstract Image denoising based on a probabilistic model of local imag e patches...	KyungHyun Cho	Unknown	2,013	{"id": "KHMdKiX2lbguE", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358345700000, "tmdate": 1358345700000, "ddate": null, "number": 66, "content": {"title": "Boltzmann Machines and Denoising Autoencoders for Image Denoising", "decision": "conferencePoster-iclr2013-workshop",...	[Review]: A brief summary of the paper's contributions, in the context of prior work. The paper proposed to use Gaussian deep Boltzmann machines (GDBM) for image denoising tasks, and it empirically compared the denoising performance to another state-of-the-art method based on stacked denoising autoencoders (Xie et al....	anonymous reviewer d5d4	null	null	{"id": "CIGoQSPKoZIKs", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362486600000, "tmdate": 1362486600000, "ddate": null, "number": 4, "content": {"title": "review of Boltzmann Machines and Denoising Autoencoders for Image Denoising", "review": "A brief summary of the paper'...	{ "criticism": 3, "example": 0, "importance_and_relevance": 5, "materials_and_methods": 17, "praise": 6, "presentation_and_reporting": 0, "results_and_discussion": 5, "suggestion_and_solution": 4, "total": 23 }	1.73913	0.160887	1.578243	1.768398	0.318966	0.029267	0.130435	0	0.217391	0.73913	0.26087	0	0.217391	0.173913	{ "criticism": 0.13043478260869565, "example": 0, "importance_and_relevance": 0.21739130434782608, "materials_and_methods": 0.7391304347826086, "praise": 0.2608695652173913, "presentation_and_reporting": 0, "results_and_discussion": 0.21739130434782608, "suggestion_and_solution": 0.17391304347826086 }	1.73913	iclr2013	openreview	null
IpmfpAGoH2KbX	Deep learning and the renormalization group	Renormalization group methods, which analyze the way in which the effective behavior of a system depends on the scale at which it is observed, are key to modern condensed-matter theory and particle physics. The aim of this paper is to compare and contrast the ideas behind the renormalization group (RG) on the one hand ...	Deep learning and the renormalization group C´edric B´eny Institut f¨ur Theoretische Physik Leibniz Universit¨at Hannover Appelstraße 2, 30167 Hannover, Germany cedric.beny@gmail.com Abstract Renormalization group (RG) methods, which model the way in which the effec- tive behavior of a system depends on the scale at wh...	Cédric Bény	Unknown	2,013	{"id": "IpmfpAGoH2KbX", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358262900000, "tmdate": 1358262900000, "ddate": null, "number": 45, "content": {"decision": "reject", "title": "Deep learning and the renormalization group", "abstract": "Renormalization group methods, which...	[Review]: Reviewer 441c, Have you taken a look at the new version of the paper? Does it go some way to addressing your concerns?	Aaron Courville	null	null	{"id": "tb0cgaJXQfgX6", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1363477320000, "tmdate": 1363477320000, "ddate": null, "number": 1, "content": {"title": "", "review": "Reviewer 441c,\r\n\r\nHave you taken a look at the new version of the paper? Does it go some way to addr...	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 2 }	0	-1.909674	1.909674	0.003153	0.027333	0.003153	0	0	0	0	0	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0, "materials_and_methods": 0, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	0	iclr2013	openreview	null
IpmfpAGoH2KbX	Deep learning and the renormalization group	Renormalization group methods, which analyze the way in which the effective behavior of a system depends on the scale at which it is observed, are key to modern condensed-matter theory and particle physics. The aim of this paper is to compare and contrast the ideas behind the renormalization group (RG) on the one hand ...	Deep learning and the renormalization group C´edric B´eny Institut f¨ur Theoretische Physik Leibniz Universit¨at Hannover Appelstraße 2, 30167 Hannover, Germany cedric.beny@gmail.com Abstract Renormalization group (RG) methods, which model the way in which the effec- tive behavior of a system depends on the scale at wh...	Cédric Bény	Unknown	2,013	{"id": "IpmfpAGoH2KbX", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358262900000, "tmdate": 1358262900000, "ddate": null, "number": 45, "content": {"decision": "reject", "title": "Deep learning and the renormalization group", "abstract": "Renormalization group methods, which...	[Review]: It is noted that the connection between RG and multi-scale modeling has been pointed out by Candes in E. J. Candès, P. Charlton and H. Helgason. Detecting highly oscillatory signals by chirplet path pursuit. Appl. Comput. Harmon. Anal. 24 14-40. where it was noted that the multi-scale basis suggested i...	Charles Martin	null	null	{"id": "rGZJRE7IJwrK3", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1392852360000, "tmdate": 1392852360000, "ddate": null, "number": 5, "content": {"title": "", "review": "It is noted that the connection between RG and multi-scale modeling has been pointed out by Candes in \r...	{ "criticism": 0, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 2, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 7 }	0.428571	-0.139596	0.568167	0.435918	0.082263	0.007346	0	0	0.142857	0.285714	0	0	0	0	{ "criticism": 0, "example": 0, "importance_and_relevance": 0.14285714285714285, "materials_and_methods": 0.2857142857142857, "praise": 0, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	0.428571	iclr2013	openreview	null
IpmfpAGoH2KbX	Deep learning and the renormalization group	Renormalization group methods, which analyze the way in which the effective behavior of a system depends on the scale at which it is observed, are key to modern condensed-matter theory and particle physics. The aim of this paper is to compare and contrast the ideas behind the renormalization group (RG) on the one hand ...	Deep learning and the renormalization group C´edric B´eny Institut f¨ur Theoretische Physik Leibniz Universit¨at Hannover Appelstraße 2, 30167 Hannover, Germany cedric.beny@gmail.com Abstract Renormalization group (RG) methods, which model the way in which the effec- tive behavior of a system depends on the scale at wh...	Cédric Bény	Unknown	2,013	{"id": "IpmfpAGoH2KbX", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358262900000, "tmdate": 1358262900000, "ddate": null, "number": 45, "content": {"decision": "reject", "title": "Deep learning and the renormalization group", "abstract": "Renormalization group methods, which...	[Review]: This paper discusses deep learning from the perspective of renormalization groups in theoretical physics. Both concepts are naturally related; however, this relation has not been formalized adequately thus far and advancing this is a novelty of the paper. The paper contains a non-technical and insightful ex...	anonymous reviewer acf4	null	null	{"id": "Qj1vSox-vpQ-U", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362219360000, "tmdate": 1362219360000, "ddate": null, "number": 3, "content": {"title": "review of Deep learning and the renormalization group", "review": "This paper discusses deep learning from the perspec...	{ "criticism": 1, "example": 0, "importance_and_relevance": 4, "materials_and_methods": 0, "praise": 2, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0, "total": 4 }	1.75	0.471623	1.278377	1.7597	0.2016	0.0097	0.25	0	1	0	0.5	0	0	0	{ "criticism": 0.25, "example": 0, "importance_and_relevance": 1, "materials_and_methods": 0, "praise": 0.5, "presentation_and_reporting": 0, "results_and_discussion": 0, "suggestion_and_solution": 0 }	1.75	iclr2013	openreview	null
IpmfpAGoH2KbX	Deep learning and the renormalization group	Renormalization group methods, which analyze the way in which the effective behavior of a system depends on the scale at which it is observed, are key to modern condensed-matter theory and particle physics. The aim of this paper is to compare and contrast the ideas behind the renormalization group (RG) on the one hand ...	Deep learning and the renormalization group C´edric B´eny Institut f¨ur Theoretische Physik Leibniz Universit¨at Hannover Appelstraße 2, 30167 Hannover, Germany cedric.beny@gmail.com Abstract Renormalization group (RG) methods, which model the way in which the effec- tive behavior of a system depends on the scale at wh...	Cédric Bény	Unknown	2,013	{"id": "IpmfpAGoH2KbX", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1358262900000, "tmdate": 1358262900000, "ddate": null, "number": 45, "content": {"decision": "reject", "title": "Deep learning and the renormalization group", "abstract": "Renormalization group methods, which...	[Review]: The model tries to relate renormalization group and deep learning, specifically hierarchical Bayesian network. The primary problems are that 1) the paper is only descriptive - it does not explain models clearly and precisely, and 2) it has no numerical experiments showing that it works. What it needs is so...	anonymous reviewer 441c	null	null	{"id": "7to37S6Q3_7Qe", "original": null, "cdate": null, "pdate": null, "odate": null, "mdate": null, "tcdate": 1362321600000, "tmdate": 1362321600000, "ddate": null, "number": 2, "content": {"title": "review of Deep learning and the renormalization group", "review": "The model tries to relate renormalization group and...	{ "criticism": 2, "example": 1, "importance_and_relevance": 1, "materials_and_methods": 7, "praise": 0, "presentation_and_reporting": 6, "results_and_discussion": 1, "suggestion_and_solution": 4, "total": 9 }	2.444444	2.191926	0.252519	2.466936	0.297227	0.022492	0.222222	0.111111	0.111111	0.777778	0	0.666667	0.111111	0.444444	{ "criticism": 0.2222222222222222, "example": 0.1111111111111111, "importance_and_relevance": 0.1111111111111111, "materials_and_methods": 0.7777777777777778, "praise": 0, "presentation_and_reporting": 0.6666666666666666, "results_and_discussion": 0.1111111111111111, "suggestion_and_solution": 0.4444444...	2.444444	iclr2013	openreview	null

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.