EE-559 Deep learning Non-volume preserving networks
|
|
- Rudolf Weaver
- 5 years ago
- Views:
Transcription
1 EE-559 Deep learning 9.4. Non-volume preserving networks François Fleuret Mon Feb 8 3:36:25 UTC 209 ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE A standard result of probability theory is that if f is continuous, invertible and [almost everywhere] differentiable, then x, p f Z x = p Z f x J f x. 3 f p Z 5 p f Z François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks / 24
2 From this equality, if f is a parametric function such that we can compute [and differentiate] p Z f x and J f x then, we can make the distribution of f Z fits the data by optimizing log p f Z x n = log p Z f x n J f x n. n n If we are able to do so, then we can synthesize a new X by sampling Z N 0, and computing f Z. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 2 / 24 If Z N 0, I, log p Z f x n = 2 f xn 2 + d log 2π. And remember that if f is is a composition of functions f = f K f we have so J f x = log J f x = K k= J f k K log J f k k= f k f x, f k f x. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 3 / 24
3 If f k are standard layers computing f z is impossible, and computing J f x is intractable. Dinh et al. 204 introduced the coupling layers to address both issues. The resulting Non-Volume Preserving network NVP is an example of a Normalizing flow Rezende and Mohamed, 205. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 4 / 24 We use here the formalism from Dinh et al Given a dimension d, a Boolean vector b {0, } d and two mappings s : R d R d t : R d R d, we define a [fully connected] coupling layer as the transformation c : R d R d x b x + b x expsb x + tb x where exp is component-wise, and is the Hadamard component-wise product. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 5 / 24
4 The expression cx = b x + b x expsb x + tb x can be understood as: forward b x unchanged, and apply to b x an invertible transformation parametrized by b x. s exp t 0 + b x cx François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 6 / 24 The consequence is that c is invertible, and if y = cx x = b y + b y tb y exp sb y. t s exp 0 + b cx x François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 7 / 24
5 The second property of this mapping is the simplicity of its Jacobian determinant. Since c i x = b i x i + b i x i exps i b x + t i b x we have, i, j, x, b i = c i x = x i and c i x j = {i=j} b i = 0 c i x = x i exps i b x + t i b x c i s i b x = {i=j} + x i x j x j }{{} 0 if b j =0 c i = {i=j} exps i b x + b j x j Hence c i x j can be non-zero only if i = j, or b i b j =. exps i b x + t i b x x j }{{} 0 if b j =0.... François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 8 / 24 If we re-order both the rows and columns of the Jacobian to put first the non-zeros entries of b, and then the zeros, it becomes lower triangular... 0 J c x = exps k x b. 0.. exps k x b its determinant remains unchanged, and we have log J f k x = s i x b i:b i =0 = i b s x b i. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 9 / 24
6 dim = 6 x = torch.empty, dim.normal_.requires_grad_ b = torch.zeros, dim b[:,:dim//2] =.0 s = nn.sequentialnn.lineardim, dim, nn.tanh t = nn.sequentialnn.lineardim, dim, nn.tanh c = b * x + - b * x * sb * x.exp + tb * x j = torch.cat[torch.autograd.gradc_k, x, retain_graph=true[0] for c_k in c[0]] printj prints tensor[[.0000, , , , , ], [ ,.0000, , , , ], [ , ,.0000, , , ], [ , , , , , ], [ , , 0.224, ,.4373, ], [ 0.030, 0.385, , , ,.0295]] François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 0 / 24 To recap, with f k, k =,..., K coupling layers, and x 0 n = x n and x k n = f k Lf = n 2 x K n f = f K f, x n k, we train by maximizing 2 + d log 2π + K log J f k k= x k n, with log J f k x = i b k s k x b k i. And to sample we just need to generate Z N 0, I and compute f Z. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks / 24
7 A coupling layer can be implemented with class NVPCouplingLayernn.Module: def init self, map_s, map_t, b: supernvpcouplinglayer, self. init self.map_s = map_s self.map_t = map_t self.b = b.clone.unsqueeze0 def forwardself, x_and_logdetjac: x, logdetjac = x_and_logdetjac s, t = self.map_sself.b * x, self.map_tself.b * x logdetjac += - self.b * s.sum y = self.b * x + - self.b * torch.exps * x + t return y, logdetjac def invertself, x: s, t = self.map_sself.b * x, self.map_tself.b * x return self.b * x + - self.b * torch.exp-s * x - t The forward here computes both the image of x and the update on the accumulated determinant of the Jacobian, i.e. x, u f x, u + J f x. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 2 / 24 We can then define a complete network with one-hidden layer tanh MLPs for the s and t mappings class NVPNetnn.Module: def init self, dim, hidden_dim, depth: supernvpnet, self. init b = torch.emptydim self.layers = nn.modulelist for d in rangedepth: if d%2 == 0: i = torch.randpermb.numel[0:b.numel // 2] b.zero_[i] = else: b = - b map_s = nn.sequentialnn.lineardim, hidden_dim, nn.tanh, nn.linearhidden_dim, dim map_t = nn.sequentialnn.lineardim, hidden_dim, nn.tanh, nn.linearhidden_dim, dim self.layers.appendnvpcouplinglayermap_s, map_t, b def forwardself, x_and_logdetjac: for m in self.layers: x_and_logdetjac = mx_and_logdetjac return x_and_logdetjac def invertself, x: for m in reversedself.layers: x = m.invertx return x François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 3 / 24
8 And the log-proba of individual samples of a batch def LogProbax_and_logdetjac: x, logdetjac = x_and_logdetjac log_p = logdetjac * x.pow2.addmath.log2 * math.pi.sum return log_p François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 4 / 24 Training is achieved by maximizing the mean log-proba batch_size = 00 model = NVPNetdim = 2, hidden_dim = 2, depth = 4 optimizer = optim.adammodel.parameters, lr = e-2 for e in rangeargs.nb_epochs: acc_loss = 0 for b in range0, nb_train_samples, batch_size: output = modelinput[b:b+batch_size], 0 loss = - LogProbaoutput.mean acc_loss = acc_loss + loss.item model.zero_grad loss.backward optimizer.step Finally, we can sample according to p f Z with z = torch.emptynb_train_samples, 2.normal_ x = model.invertz François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 5 / 24
9 4 Real Synth 4 Real Synth Real Synth 4 Real Synth François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 6 / 24 Dinh et al. 206 apply this approach to convolutional layers by using bs Published as a conference paper at ICLR 207 consistent with the activation map structure, and reducing the map size while increasing the number of channels Figure 3: Masking schemes for affine coupling layers. On the left, a spatial checkerboard pattern mask. On the right, a channel-wise masking. The squeezing operation reduces the 4 4 tensor on the left into a tensor on the right. Before the squeezing operation, a checkerboard pattern is used for coupling layers while a channel-wise masking pattern is used afterward. see Figure 2b, { y:d = x :d y d+:d = x d+:d exp sx :d + tx :d Dinh et al., { x:d = y :d x d+:d = y d+:d ty :d exp sy :d, 8 meaning that sampling is as efficient as inference for this model. Note again that computing the inverse of the coupling layer does not require computing the inverse of s or t, so these functions can François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 7 / 24
10 Published as a conference paper at ICLR 207 They combine these layers by alternating masks, and branching out half of the channels at certain points to forward them unchanged. + x = + x z z2 z3 z4 f 3 2 z3 h4 f 2 = + x = z z2 h3 h4 x x2 x3 x4 f a In this alternating pattern, units which remain identical in one transformation are modified in the next. b Factoring out variables. At each step, half the variables are directly modeled as Gaussians, while the other half undergo further transformation. 3.6 Multi-scale architecture Figure 4: Composition schemes for affine coupling layers. Dinh et al., 206 We implement a multi-scale architecture using a squeezing operation: for each channel, it divides the image into subsquares of shape 2 2 c, then reshapes them into subsquares of shape 4c. The squeezing operation transforms an s s c tensor into an s 2 s 2 4c tensor see Figure 3, effectively trading spatial size for number of channels. At each scale, we combine several operations into a sequence: we first apply three coupling layers with alternating checkerboard masks, then perform a squeezing operation, and finally apply three more coupling layers with alternating channel-wise masking. The channel-wise masking is chosen so that the resulting partitioning is not redundant with the previous checkerboard masking see Figure 3. For the final scale, we only apply four coupling layers with alternating checkerboard masks. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 8 / 24 Propagating a D dimensional vector through all the coupling layers would be cumbersome, in terms of computational and memory cost, and in terms of the number of parameters that would need to be trained. For this reason we follow the design choice of [57] and factor out half of the dimensions at regular intervals see Equation 4. We can define this operation recursively see Figure 4b, The structure for generating images consists of 2 stages h 0 = x 3 z i+, h i+ = f i+ h i 4 z L = f L h L 5 z = z,..., z L. 6 In our experiments, 3 checkerboard we use this coupling operationlayers, for i < L. The sequence of coupling-squeezing-coupling operations a squeezing described above layer, is performed per layer when computing f i Equation 4. At each layer, as 3 the spatial channel resolution coupling is reduced, layers, the number of hidden layer features in s and t is doubled. All variables a factor-out which have layer. been factored out at different scales are concatenated to obtain the final transformed output Equation 6. stage As a consequence, the model must Gaussianize units which are factored out at a finer scale in an earlier layer 4 before checkerboard those which coupling are factored layers out at a coarser scale in a later layer. This results in the definition aof factor-out intermediary layer. levels of representation [53, 49] corresponding to more local, fine-grained features as shown in Appendix D. Moreover, Gaussianizing and factoring out units in earlier layers has the practical benefit of distributing the loss function throughout the network, following the philosophy similar to guiding intermediate layers using intermediate classifiers [40]. It also reduces significantly the amount of computation and memory used by the model, allowing us to train larger models. The s and t mappings get more complex in the later layers. 6 François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 9 / 24
11 Figure 7: Samples from a model trained on Imagenet Dinh et al., 206 François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 20 / 24 3 Figure 8: Samples from a model trained on CelebA. Dinh et al., 206 François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 2 / 24
12 Figure 9: Samples from a model trained on LSUN bedroom category. Dinh et al., 206 François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 22 / 24 5 Figure 0: Samples from a model trained on LSUN church outdoor category. Dinh et al., 206 François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 23 / 24
13 Figure 6: Manifold generated from four examples in the dataset. Clockwise from top left: CelebA, Imagenet 64 64, LSUN tower, LSUN bedroom. over sample quality in a limited capacity setting. As a result, our model outputs sometimes highly improbable samples as we can notice especially on CelebA. As opposed to variational Dinh etautoencoders, al., 206 the samples generated from our model look not only globally coherent but also sharp. Our hypothesis is that as opposed to these models, real NVP does not rely on fixed form reconstruction cost like an L 2 norm which tends to reward capturing low frequency components more heavily than high frequency components. Unlike autoregressive models, sampling from our model is done very efficiently as it is parallelized over input dimensions. On Imagenet and LSUN, our model seems to have captured well the notion of background/foreground and lighting interactions such as luminosity and consistent light source direction for reflectance and shadows. François Fleuret EE-559 Deep learning / 9.4. Non-volume preserving networks 24 / 24 We also illustrate the smooth semantically consistent meaning of our latent variables. In the latent space, we define a manifold based on four validation examples z, z 2, z 3, z 4, and parametrized by two parameters φ and φ by, z = cosφ cosφ z + sinφ z 2 + sinφ cosφ z 3 + sinφ z 4. 9 We project the resulting manifold back into the data space by computing gz. Results are shown Figure 6. We observe that the model seems to have organized the latent space with a notion of meaning that goes well beyond pixel space interpolation. More visualization are shown in the Appendix. References 5 Discussion and conclusion L. InDinh, this paper, D. Krueger, we haveand defined Y. Bengio. a class ofnice: invertible non-linear functions independent with tractable components Jacobian determinant, estimation. enabling CoRR, exact abs/40.856, and tractable log-likelihood 204. evaluation, inference, and sampling. We have shown that this class of generative model achieves competitive performances, both in terms of sample quality L. anddinh, log-likelihood. J. Sohl-Dickstein, Many avenues and exist S. Bengio. to further Density improveestimation the functional using formreal of the NVP. transformations, CoRR, forabs/ , instance by exploiting 206. the latest advances in dilated convolutions [69] and residual networks D. architectures Rezende and [60]. S. Mohamed. Variational inference with normalizing flows. CoRR, This abs/ , paper presented205. a technique bridging the gap between auto-regressive models, variational autoencoders, and generative adversarial networks. Like auto-regressive models, it allows tractable and exact log-likelihood evaluation for training. It allows however a much more flexible functional form, similar to that in the generative model of variational autoencoders. This allows for fast and exact sampling from the model distribution. Like GANs, and unlike variational autoencoders, our technique does not require the use of a fixed form reconstruction cost, and instead defines a cost in terms of higher level features, generating sharper images. Finally, unlike both variational autoencoders and GANs, our technique is able to learn a semantically meaningful latent space which is as high dimensional as the input space. This may make the algorithm particularly well suited to semi-supervised learning tasks, as we hope to explore in future work. 9
Deep Generative Models Variational Autoencoders
Deep Generative Models Variational Autoencoders Sudeshna Sarkar 5 April 2017 Generative Nets Generative models that represent probability distributions over multiple variables in some way. Directed Generative
More informationEE-559 Deep learning Networks for semantic segmentation
EE-559 Deep learning 7.4. Networks for semantic segmentation François Fleuret https://fleuret.org/ee559/ Mon Feb 8 3:35:5 UTC 209 ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE The historical approach to image
More informationDEEP LEARNING PART THREE - DEEP GENERATIVE MODELS CS/CNS/EE MACHINE LEARNING & DATA MINING - LECTURE 17
DEEP LEARNING PART THREE - DEEP GENERATIVE MODELS CS/CNS/EE 155 - MACHINE LEARNING & DATA MINING - LECTURE 17 GENERATIVE MODELS DATA 3 DATA 4 example 1 DATA 5 example 2 DATA 6 example 3 DATA 7 number of
More informationGenerative Adversarial Networks (GANs) Based on slides from Ian Goodfellow s NIPS 2016 tutorial
Generative Adversarial Networks (GANs) Based on slides from Ian Goodfellow s NIPS 2016 tutorial Generative Modeling Density estimation Sample generation Training examples Model samples Next Video Frame
More information(University Improving of Montreal) Generative Adversarial Networks with Denoising Feature Matching / 17
Improving Generative Adversarial Networks with Denoising Feature Matching David Warde-Farley 1 Yoshua Bengio 1 1 University of Montreal, ICLR,2017 Presenter: Bargav Jayaraman Outline 1 Introduction 2 Background
More informationDay 3 Lecture 1. Unsupervised Learning
Day 3 Lecture 1 Unsupervised Learning Semi-supervised and transfer learning Myth: you can t do deep learning unless you have a million labelled examples for your problem. Reality You can learn useful representations
More informationNeural Networks and Deep Learning
Neural Networks and Deep Learning Example Learning Problem Example Learning Problem Celebrity Faces in the Wild Machine Learning Pipeline Raw data Feature extract. Feature computation Inference: prediction,
More informationDeep Generative Image Models using a Laplacian Pyramid of Adversarial Networks Supplementary Material
Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks Supplementary Material Emily Denton Dept. of Computer Science Courant Institute New York University Soumith Chintala Arthur
More informationAdversarially Learned Inference
Institut des algorithmes d apprentissage de Montréal Adversarially Learned Inference Aaron Courville CIFAR Fellow Université de Montréal Joint work with: Vincent Dumoulin, Ishmael Belghazi, Olivier Mastropietro,
More informationGenerative Adversarial Networks (GANs) Ian Goodfellow, Research Scientist MLSLP Keynote, San Francisco
Generative Adversarial Networks (GANs) Ian Goodfellow, Research Scientist MLSLP Keynote, San Francisco 2016-09-13 Generative Modeling Density estimation Sample generation Training examples Model samples
More informationGenerative Adversarial Networks (GANs)
Generative Adversarial Networks (GANs) Hossein Azizpour Most of the slides are courtesy of Dr. Ian Goodfellow (Research Scientist at OpenAI) and from his presentation at NIPS 2016 tutorial Note. I am generally
More informationAutoencoders. Stephen Scott. Introduction. Basic Idea. Stacked AE. Denoising AE. Sparse AE. Contractive AE. Variational AE GAN.
Stacked Denoising Sparse Variational (Adapted from Paul Quint and Ian Goodfellow) Stacked Denoising Sparse Variational Autoencoding is training a network to replicate its input to its output Applications:
More informationarxiv: v1 [cs.cv] 17 Nov 2016
Inverting The Generator Of A Generative Adversarial Network arxiv:1611.05644v1 [cs.cv] 17 Nov 2016 Antonia Creswell BICV Group Bioengineering Imperial College London ac2211@ic.ac.uk Abstract Anil Anthony
More informationDiffusion Wavelets for Natural Image Analysis
Diffusion Wavelets for Natural Image Analysis Tyrus Berry December 16, 2011 Contents 1 Project Description 2 2 Introduction to Diffusion Wavelets 2 2.1 Diffusion Multiresolution............................
More informationPixel-level Generative Model
Pixel-level Generative Model Generative Image Modeling Using Spatial LSTMs (2015NIPS) L. Theis and M. Bethge University of Tübingen, Germany Pixel Recurrent Neural Networks (2016ICML) A. van den Oord,
More informationAuto-Encoding Variational Bayes
Auto-Encoding Variational Bayes Diederik P (Durk) Kingma, Max Welling University of Amsterdam Ph.D. Candidate, advised by Max Durk Kingma D.P. Kingma Max Welling Problem class Directed graphical model:
More informationImage Restoration with Deep Generative Models
Image Restoration with Deep Generative Models Raymond A. Yeh *, Teck-Yian Lim *, Chen Chen, Alexander G. Schwing, Mark Hasegawa-Johnson, Minh N. Do Department of Electrical and Computer Engineering, University
More informationStochastic Simulation with Generative Adversarial Networks
Stochastic Simulation with Generative Adversarial Networks Lukas Mosser, Olivier Dubrule, Martin J. Blunt lukas.mosser15@imperial.ac.uk, o.dubrule@imperial.ac.uk, m.blunt@imperial.ac.uk (Deep) Generative
More informationOne Network to Solve Them All Solving Linear Inverse Problems using Deep Projection Models
One Network to Solve Them All Solving Linear Inverse Problems using Deep Projection Models [Supplemental Materials] 1. Network Architecture b ref b ref +1 We now describe the architecture of the networks
More informationImage Processing. Filtering. Slide 1
Image Processing Filtering Slide 1 Preliminary Image generation Original Noise Image restoration Result Slide 2 Preliminary Classic application: denoising However: Denoising is much more than a simple
More informationSemantic Segmentation. Zhongang Qi
Semantic Segmentation Zhongang Qi qiz@oregonstate.edu Semantic Segmentation "Two men riding on a bike in front of a building on the road. And there is a car." Idea: recognizing, understanding what's in
More informationDeep Learning With Noise
Deep Learning With Noise Yixin Luo Computer Science Department Carnegie Mellon University yixinluo@cs.cmu.edu Fan Yang Department of Mathematical Sciences Carnegie Mellon University fanyang1@andrew.cmu.edu
More informationNatural Language Processing CS 6320 Lecture 6 Neural Language Models. Instructor: Sanda Harabagiu
Natural Language Processing CS 6320 Lecture 6 Neural Language Models Instructor: Sanda Harabagiu In this lecture We shall cover: Deep Neural Models for Natural Language Processing Introduce Feed Forward
More informationLearning and Recognizing Visual Object Categories Without First Detecting Features
Learning and Recognizing Visual Object Categories Without First Detecting Features Daniel Huttenlocher 2007 Joint work with D. Crandall and P. Felzenszwalb Object Category Recognition Generic classes rather
More informationConvolutional Neural Networks
Lecturer: Barnabas Poczos Introduction to Machine Learning (Lecture Notes) Convolutional Neural Networks Disclaimer: These notes have not been subjected to the usual scrutiny reserved for formal publications.
More informationUnsupervised Learning
Deep Learning for Graphics Unsupervised Learning Niloy Mitra Iasonas Kokkinos Paul Guerrero Vladimir Kim Kostas Rematas Tobias Ritschel UCL UCL/Facebook UCL Adobe Research U Washington UCL Timetable Niloy
More informationLecture 10 CNNs on Graphs
Lecture 10 CNNs on Graphs CMSC 35246: Deep Learning Shubhendu Trivedi & Risi Kondor University of Chicago April 26, 2017 Two Scenarios For CNNs on graphs, we have two distinct scenarios: Scenario 1: Each
More informationScaled representations
Scaled representations Big bars (resp. spots, hands, etc.) and little bars are both interesting Stripes and hairs, say Inefficient to detect big bars with big filters And there is superfluous detail in
More informationHidden Units. Sargur N. Srihari
Hidden Units Sargur N. srihari@cedar.buffalo.edu 1 Topics in Deep Feedforward Networks Overview 1. Example: Learning XOR 2. Gradient-Based Learning 3. Hidden Units 4. Architecture Design 5. Backpropagation
More informationAlternatives to Direct Supervision
CreativeAI: Deep Learning for Graphics Alternatives to Direct Supervision Niloy Mitra Iasonas Kokkinos Paul Guerrero Nils Thuerey Tobias Ritschel UCL UCL UCL TUM UCL Timetable Theory and Basics State of
More informationGenerative Models in Deep Learning. Sargur N. Srihari
Generative Models in Deep Learning Sargur N. Srihari srihari@cedar.buffalo.edu 1 Topics 1. Need for Probabilities in Machine Learning 2. Representations 1. Generative and Discriminative Models 2. Directed/Undirected
More information3.1. Solution for white Gaussian noise
Low complexity M-hypotheses detection: M vectors case Mohammed Nae and Ahmed H. Tewk Dept. of Electrical Engineering University of Minnesota, Minneapolis, MN 55455 mnae,tewk@ece.umn.edu Abstract Low complexity
More informationGENERATIVE ADVERSARIAL NETWORKS (GAN) Presented by Omer Stein and Moran Rubin
GENERATIVE ADVERSARIAL NETWORKS (GAN) Presented by Omer Stein and Moran Rubin GENERATIVE MODEL Given a training dataset, x, try to estimate the distribution, Pdata(x) Explicitly or Implicitly (GAN) Explicitly
More informationCMU Lecture 18: Deep learning and Vision: Convolutional neural networks. Teacher: Gianni A. Di Caro
CMU 15-781 Lecture 18: Deep learning and Vision: Convolutional neural networks Teacher: Gianni A. Di Caro DEEP, SHALLOW, CONNECTED, SPARSE? Fully connected multi-layer feed-forward perceptrons: More powerful
More informationImage Transformation via Neural Network Inversion
Image Transformation via Neural Network Inversion Asha Anoosheh Rishi Kapadia Jared Rulison Abstract While prior experiments have shown it is possible to approximately reconstruct inputs to a neural net
More informationLecture 5 2D Transformation
Lecture 5 2D Transformation What is a transformation? In computer graphics an object can be transformed according to position, orientation and size. Exactly what it says - an operation that transforms
More informationReplacing Neural Networks with Black-Box ODE Solvers
Replacing Neural Networks with Black-Box ODE Solvers Tian Qi Chen, Yulia Rubanova, Jesse Bettencourt, David Duvenaud University of Toronto, Vector Institute Resnets are Euler integrators Middle layers
More informationNeural Networks: promises of current research
April 2008 www.apstat.com Current research on deep architectures A few labs are currently researching deep neural network training: Geoffrey Hinton s lab at U.Toronto Yann LeCun s lab at NYU Our LISA lab
More informationBayesian model ensembling using meta-trained recurrent neural networks
Bayesian model ensembling using meta-trained recurrent neural networks Luca Ambrogioni l.ambrogioni@donders.ru.nl Umut Güçlü u.guclu@donders.ru.nl Yağmur Güçlütürk y.gucluturk@donders.ru.nl Julia Berezutskaya
More informationLearning visual odometry with a convolutional network
Learning visual odometry with a convolutional network Kishore Konda 1, Roland Memisevic 2 1 Goethe University Frankfurt 2 University of Montreal konda.kishorereddy@gmail.com, roland.memisevic@gmail.com
More informationCS 4495 Computer Vision. Linear Filtering 2: Templates, Edges. Aaron Bobick. School of Interactive Computing. Templates/Edges
CS 4495 Computer Vision Linear Filtering 2: Templates, Edges Aaron Bobick School of Interactive Computing Last time: Convolution Convolution: Flip the filter in both dimensions (right to left, bottom to
More informationLecture 21 : A Hybrid: Deep Learning and Graphical Models
10-708: Probabilistic Graphical Models, Spring 2018 Lecture 21 : A Hybrid: Deep Learning and Graphical Models Lecturer: Kayhan Batmanghelich Scribes: Paul Liang, Anirudha Rayasam 1 Introduction and Motivation
More informationAuxiliary Guided Autoregressive Variational Autoencoders
Auxiliary Guided Autoregressive Variational Autoencoders Thomas Lucas, Jakob Verbeek To cite this version: Thomas Lucas, Jakob Verbeek. Auxiliary Guided Autoregressive Variational Autoencoders. 2017.
More informationComputer Vision 2. SS 18 Dr. Benjamin Guthier Professur für Bildverarbeitung. Computer Vision 2 Dr. Benjamin Guthier
Computer Vision 2 SS 18 Dr. Benjamin Guthier Professur für Bildverarbeitung Computer Vision 2 Dr. Benjamin Guthier 1. IMAGE PROCESSING Computer Vision 2 Dr. Benjamin Guthier Content of this Chapter Non-linear
More informationImplicit generative models: dual vs. primal approaches
Implicit generative models: dual vs. primal approaches Ilya Tolstikhin MPI for Intelligent Systems ilya@tue.mpg.de Machine Learning Summer School 2017 Tübingen, Germany Contents 1. Unsupervised generative
More informationResearch on Pruning Convolutional Neural Network, Autoencoder and Capsule Network
Research on Pruning Convolutional Neural Network, Autoencoder and Capsule Network Tianyu Wang Australia National University, Colledge of Engineering and Computer Science u@anu.edu.au Abstract. Some tasks,
More informationModel Generalization and the Bias-Variance Trade-Off
Charu C. Aggarwal IBM T J Watson Research Center Yorktown Heights, NY Model Generalization and the Bias-Variance Trade-Off Neural Networks and Deep Learning, Springer, 2018 Chapter 4, Section 4.1-4.2 What
More informationKnow your data - many types of networks
Architectures Know your data - many types of networks Fixed length representation Variable length representation Online video sequences, or samples of different sizes Images Specific architectures for
More informationEE795: Computer Vision and Intelligent Systems
EE795: Computer Vision and Intelligent Systems Spring 2012 TTh 17:30-18:45 WRI C225 Lecture 04 130131 http://www.ee.unlv.edu/~b1morris/ecg795/ 2 Outline Review Histogram Equalization Image Filtering Linear
More informationWhat is Multigrid? They have been extended to solve a wide variety of other problems, linear and nonlinear.
AMSC 600/CMSC 760 Fall 2007 Solution of Sparse Linear Systems Multigrid, Part 1 Dianne P. O Leary c 2006, 2007 What is Multigrid? Originally, multigrid algorithms were proposed as an iterative method to
More informationLecture 19: Generative Adversarial Networks
Lecture 19: Generative Adversarial Networks Roger Grosse 1 Introduction Generative modeling is a type of machine learning where the aim is to model the distribution that a given set of data (e.g. images,
More information08 An Introduction to Dense Continuous Robotic Mapping
NAVARCH/EECS 568, ROB 530 - Winter 2018 08 An Introduction to Dense Continuous Robotic Mapping Maani Ghaffari March 14, 2018 Previously: Occupancy Grid Maps Pose SLAM graph and its associated dense occupancy
More informationConvolution Neural Networks for Chinese Handwriting Recognition
Convolution Neural Networks for Chinese Handwriting Recognition Xu Chen Stanford University 450 Serra Mall, Stanford, CA 94305 xchen91@stanford.edu Abstract Convolutional neural networks have been proven
More informationFrom processing to learning on graphs
From processing to learning on graphs Patrick Pérez Maths and Images in Paris IHP, 2 March 2017 Signals on graphs Natural graph: mesh, network, etc., related to a real structure, various signals can live
More informationGenerative Adversarial Text to Image Synthesis
Generative Adversarial Text to Image Synthesis Scott Reed, Zeynep Akata, Xinchen Yan, Lajanugen Logeswaran, Bernt Schiele, Honglak Lee Presented by: Jingyao Zhan Contents Introduction Related Work Method
More informationQuantitative Evaluation of Generative Adversarial Networks and Improved Training Techniques
Quantitative Evaluation of Generative Adversarial Networks and Improved Training Techniques by Yadong Li to obtain the degree of Master of Science at the Delft University of Technology, to be defended
More informationGAN Frontiers/Related Methods
GAN Frontiers/Related Methods Improving GAN Training Improved Techniques for Training GANs (Salimans, et. al 2016) CSC 2541 (07/10/2016) Robin Swanson (robin@cs.toronto.edu) Training GANs is Difficult
More informationMachine Learning for Physicists Lecture 6. Summer 2017 University of Erlangen-Nuremberg Florian Marquardt
Machine Learning for Physicists Lecture 6 Summer 2017 University of Erlangen-Nuremberg Florian Marquardt Channels MxM image MxM image K K 3 channels conv 6 channels in any output channel, each pixel receives
More informationNatural Language Processing with Deep Learning CS224N/Ling284. Christopher Manning Lecture 4: Backpropagation and computation graphs
Natural Language Processing with Deep Learning CS4N/Ling84 Christopher Manning Lecture 4: Backpropagation and computation graphs Lecture Plan Lecture 4: Backpropagation and computation graphs 1. Matrix
More informationGrundlagen der Künstlichen Intelligenz
Grundlagen der Künstlichen Intelligenz Unsupervised learning Daniel Hennes 29.01.2018 (WS 2017/18) University Stuttgart - IPVS - Machine Learning & Robotics 1 Today Supervised learning Regression (linear
More informationKnowledge Discovery and Data Mining 1 (VO) ( )
Knowledge Discovery and Data Mining 1 (VO) (707.003) Data Matrices and Vector Space Model Denis Helic KTI, TU Graz Nov 6, 2014 Denis Helic (KTI, TU Graz) KDDM1 Nov 6, 2014 1 / 55 Big picture: KDDM Probability
More informationEnergy Based Models, Restricted Boltzmann Machines and Deep Networks. Jesse Eickholt
Energy Based Models, Restricted Boltzmann Machines and Deep Networks Jesse Eickholt ???? Who s heard of Energy Based Models (EBMs) Restricted Boltzmann Machines (RBMs) Deep Belief Networks Auto-encoders
More informationInverting The Generator Of A Generative Adversarial Network
1 Inverting The Generator Of A Generative Adversarial Network Antonia Creswell and Anil A Bharath, Imperial College London arxiv:1802.05701v1 [cs.cv] 15 Feb 2018 Abstract Generative adversarial networks
More informationELEC Dr Reji Mathew Electrical Engineering UNSW
ELEC 4622 Dr Reji Mathew Electrical Engineering UNSW Review of Motion Modelling and Estimation Introduction to Motion Modelling & Estimation Forward Motion Backward Motion Block Motion Estimation Motion
More informationProbabilistic Graphical Models
School of Computer Science Probabilistic Graphical Models Theory of Variational Inference: Inner and Outer Approximation Eric Xing Lecture 14, February 29, 2016 Reading: W & J Book Chapters Eric Xing @
More informationIn-Place Activated BatchNorm for Memory- Optimized Training of DNNs
In-Place Activated BatchNorm for Memory- Optimized Training of DNNs Samuel Rota Bulò, Lorenzo Porzi, Peter Kontschieder Mapillary Research Paper: https://arxiv.org/abs/1712.02616 Code: https://github.com/mapillary/inplace_abn
More informationInception Network Overview. David White CS793
Inception Network Overview David White CS793 So, Leonardo DiCaprio dreams about dreaming... https://m.media-amazon.com/images/m/mv5bmjaxmzy3njcxnf5bml5banbnxkftztcwnti5otm0mw@@._v1_sy1000_cr0,0,675,1 000_AL_.jpg
More informationErasing Haar Coefficients
Recap Haar simple and fast wavelet transform Limitations not smooth enough: blocky How to improve? classical approach: basis functions Lifting: transforms 1 Erasing Haar Coefficients 2 Page 1 Classical
More informationDropout. Sargur N. Srihari This is part of lecture slides on Deep Learning:
Dropout Sargur N. srihari@buffalo.edu This is part of lecture slides on Deep Learning: http://www.cedar.buffalo.edu/~srihari/cse676 1 Regularization Strategies 1. Parameter Norm Penalties 2. Norm Penalties
More informationSharpening through spatial filtering
Sharpening through spatial filtering Stefano Ferrari Università degli Studi di Milano stefano.ferrari@unimi.it Methods for Image Processing academic year 2017 2018 Sharpening The term sharpening is referred
More informationarxiv: v4 [stat.ml] 4 Dec 2017
On the Effects of Batch and Weight Normalization in Generative Adversarial Networks arxiv:1704.03971v4 [stat.ml] 4 Dec 2017 Abstract Sitao Xiang 1 1, 2, 3 Hao Li 1 University of Southern California 2 Pinscreen
More informationSoftware Testing. 1. Testing is the process of demonstrating that errors are not present.
What is Testing? Software Testing Many people understand many definitions of testing :. Testing is the process of demonstrating that errors are not present.. The purpose of testing is to show that a program
More informationAmortised MAP Inference for Image Super-resolution. Casper Kaae Sønderby, Jose Caballero, Lucas Theis, Wenzhe Shi & Ferenc Huszár ICLR 2017
Amortised MAP Inference for Image Super-resolution Casper Kaae Sønderby, Jose Caballero, Lucas Theis, Wenzhe Shi & Ferenc Huszár ICLR 2017 Super Resolution Inverse problem: Given low resolution representation
More informationMode Regularized Generative Adversarial Networks
Mode Regularized Generative Adversarial Networks Tong Che 1 Yanran Li 2 Athul Paul Jacob 3 Yoshua Bengio 1 Wenjie Li 2 1 Montreal Institute for Learning Algorithms, Universite de Montreal, Montreal, Canada
More informationCS325 Artificial Intelligence Ch. 20 Unsupervised Machine Learning
CS325 Artificial Intelligence Cengiz Spring 2013 Unsupervised Learning Missing teacher No labels, y Just input data, x What can you learn with it? Unsupervised Learning Missing teacher No labels, y Just
More informationSpatial and multi-scale data assimilation in EO-LDAS. Technical Note for EO-LDAS project/nceo. P. Lewis, UCL NERC NCEO
Spatial and multi-scale data assimilation in EO-LDAS Technical Note for EO-LDAS project/nceo P. Lewis, UCL NERC NCEO Abstract Email: p.lewis@ucl.ac.uk 2 May 2012 In this technical note, spatial data assimilation
More informationGenerative Adversarial Network
Generative Adversarial Network Many slides from NIPS 2014 Ian J. Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, Yoshua Bengio Generative adversarial
More informationLecture 7: Semantic Segmentation
Semantic Segmentation CSED703R: Deep Learning for Visual Recognition (207F) Segmenting images based on its semantic notion Lecture 7: Semantic Segmentation Bohyung Han Computer Vision Lab. bhhanpostech.ac.kr
More informationCSE 250B Project Assignment 4
CSE 250B Project Assignment 4 Hani Altwary haltwa@cs.ucsd.edu Kuen-Han Lin kul016@ucsd.edu Toshiro Yamada toyamada@ucsd.edu Abstract The goal of this project is to implement the Semi-Supervised Recursive
More informationMarkov Random Fields for Recognizing Textures modeled by Feature Vectors
Markov Random Fields for Recognizing Textures modeled by Feature Vectors Juliette Blanchet, Florence Forbes, and Cordelia Schmid Inria Rhône-Alpes, 655 avenue de l Europe, Montbonnot, 38334 Saint Ismier
More information27: Hybrid Graphical Models and Neural Networks
10-708: Probabilistic Graphical Models 10-708 Spring 2016 27: Hybrid Graphical Models and Neural Networks Lecturer: Matt Gormley Scribes: Jakob Bauer Otilia Stretcu Rohan Varma 1 Motivation We first look
More information3. Data Structures for Image Analysis L AK S H M O U. E D U
3. Data Structures for Image Analysis L AK S H M AN @ O U. E D U Different formulations Can be advantageous to treat a spatial grid as a: Levelset Matrix Markov chain Topographic map Relational structure
More informationConditional Random Fields as Recurrent Neural Networks
BIL722 - Deep Learning for Computer Vision Conditional Random Fields as Recurrent Neural Networks S. Zheng, S. Jayasumana, B. Romera-Paredes V. Vineet, Z. Su, D. Du, C. Huang, P.H.S. Torr Introduction
More informationMachine Learning. Deep Learning. Eric Xing (and Pengtao Xie) , Fall Lecture 8, October 6, Eric CMU,
Machine Learning 10-701, Fall 2015 Deep Learning Eric Xing (and Pengtao Xie) Lecture 8, October 6, 2015 Eric Xing @ CMU, 2015 1 A perennial challenge in computer vision: feature engineering SIFT Spin image
More informationIntroduction to GAN. Generative Adversarial Networks. Junheng(Jeff) Hao
Introduction to GAN Generative Adversarial Networks Junheng(Jeff) Hao Adversarial Training is the coolest thing since sliced bread. -- Yann LeCun Roadmap 1. Generative Modeling 2. GAN 101: What is GAN?
More informationSo we have been talking about 3D viewing, the transformations pertaining to 3D viewing. Today we will continue on it. (Refer Slide Time: 1:15)
Introduction to Computer Graphics Dr. Prem Kalra Department of Computer Science and Engineering Indian Institute of Technology, Delhi Lecture - 8 3D Viewing So we have been talking about 3D viewing, the
More informationVariational Autoencoders. Sargur N. Srihari
Variational Autoencoders Sargur N. srihari@cedar.buffalo.edu Topics 1. Generative Model 2. Standard Autoencoder 3. Variational autoencoders (VAE) 2 Generative Model A variational autoencoder (VAE) is a
More informationMachine Learning. The Breadth of ML Neural Networks & Deep Learning. Marc Toussaint. Duy Nguyen-Tuong. University of Stuttgart
Machine Learning The Breadth of ML Neural Networks & Deep Learning Marc Toussaint University of Stuttgart Duy Nguyen-Tuong Bosch Center for Artificial Intelligence Summer 2017 Neural Networks Consider
More information2D and 3D Transformations AUI Course Denbigh Starkey
2D and 3D Transformations AUI Course Denbigh Starkey. Introduction 2 2. 2D transformations using Cartesian coordinates 3 2. Translation 3 2.2 Rotation 4 2.3 Scaling 6 3. Introduction to homogeneous coordinates
More informationImage processing in frequency Domain
Image processing in frequency Domain Introduction to Frequency Domain Deal with images in: -Spatial domain -Frequency domain Frequency Domain In the frequency or Fourier domain, the value and location
More informationCHAPTER 6 Parametric Spline Curves
CHAPTER 6 Parametric Spline Curves When we introduced splines in Chapter 1 we focused on spline curves, or more precisely, vector valued spline functions. In Chapters 2 and 4 we then established the basic
More informationDescribable Visual Attributes for Face Verification and Image Search
Advanced Topics in Multimedia Analysis and Indexing, Spring 2011, NTU. 1 Describable Visual Attributes for Face Verification and Image Search Kumar, Berg, Belhumeur, Nayar. PAMI, 2011. Ryan Lei 2011/05/05
More informationDeep Learning. Vladimir Golkov Technical University of Munich Computer Vision Group
Deep Learning Vladimir Golkov Technical University of Munich Computer Vision Group 1D Input, 1D Output target input 2 2D Input, 1D Output: Data Distribution Complexity Imagine many dimensions (data occupies
More informationCOMP 551 Applied Machine Learning Lecture 16: Deep Learning
COMP 551 Applied Machine Learning Lecture 16: Deep Learning Instructor: Ryan Lowe (ryan.lowe@cs.mcgill.ca) Slides mostly by: Class web page: www.cs.mcgill.ca/~hvanho2/comp551 Unless otherwise noted, all
More informationCHAPTER 5 GLOBAL AND LOCAL FEATURES FOR FACE RECOGNITION
122 CHAPTER 5 GLOBAL AND LOCAL FEATURES FOR FACE RECOGNITION 5.1 INTRODUCTION Face recognition, means checking for the presence of a face from a database that contains many faces and could be performed
More informationPARTIAL STYLE TRANSFER USING WEAKLY SUPERVISED SEMANTIC SEGMENTATION. Shin Matsuo Wataru Shimoda Keiji Yanai
PARTIAL STYLE TRANSFER USING WEAKLY SUPERVISED SEMANTIC SEGMENTATION Shin Matsuo Wataru Shimoda Keiji Yanai Department of Informatics, The University of Electro-Communications, Tokyo 1-5-1 Chofugaoka,
More informationUnsupervised Learning of Spatiotemporally Coherent Metrics
Unsupervised Learning of Spatiotemporally Coherent Metrics Ross Goroshin, Joan Bruna, Jonathan Tompson, David Eigen, Yann LeCun arxiv 2015. Presented by Jackie Chu Contributions Insight between slow feature
More informationDeep Boltzmann Machines
Deep Boltzmann Machines Sargur N. Srihari srihari@cedar.buffalo.edu Topics 1. Boltzmann machines 2. Restricted Boltzmann machines 3. Deep Belief Networks 4. Deep Boltzmann machines 5. Boltzmann machines
More informationScientific Visualization Example exam questions with commented answers
Scientific Visualization Example exam questions with commented answers The theoretical part of this course is evaluated by means of a multiple- choice exam. The questions cover the material mentioned during
More informationAutoencoder. Representation learning (related to dictionary learning) Both the input and the output are x
Deep Learning 4 Autoencoder, Attention (spatial transformer), Multi-modal learning, Neural Turing Machine, Memory Networks, Generative Adversarial Net Jian Li IIIS, Tsinghua Autoencoder Autoencoder Unsupervised
More information