Singing Voice Separation using Generative Adversarial Networks

Size: px
Start display at page:

Download "Singing Voice Separation using Generative Adversarial Networks"

Transcription

1 Singing Voice Separation using Generative Adversarial Networks Hyeong-seok Choi, Kyogu Lee Music and Audio Research Group Graduate School of Convergence Science and Technology Seoul National University {kekepa15, Ju-heon Lee College of Liberal Studies Seoul National University Abstract In this paper, we propose a novel approach extending Wasserstein generative adversarial networks (GANs) [3] to separate singing voice from the mixture signal. We used the mixture signal as a condition to generate singing voices and applied the U-net style network for the stable training of the model. Experiments with the DSD100 dataset show the promising results with the potential of using the GANs for music source separation. 1 Introduction Music source separation is the process of separating a specific source from a music signal. Separating the source from the mixture signal can be interpreted as maximizing the likelihood of the source from a given mixture. Our task is to perform this task using GANs [1] which are classified as a method to maximize the likelihood by using implicit density. GANs are usually used to produce samples from noise, but in recent years, research [7,8] has been under way to better tailor the desired sample with a certain constraint. In this paper, our research aims to generate singing voice signals using mixture signals as a condition. Background GANs are generative model that learns a function generator G θ to map noise samples z p(z) into the real data space. The main idea of training GANs is often described as a mini-max game between two players which are discriminator (D) and generator (G) [1]. The input of D is either real sample x P r or fake sample x x x P g and the mission of D is to classify x x x as fake and to classify x as real. Many improved GANs model was attempted [,3,5] and one of the notable GANs studies that provides both theoretic background and practical result is the Wasserstein GANs. It is a model that tries to reduce the Wasserstein distance between the data distribution (P r ) and the generated sample distribution (P g ). Using the Wasserstein distance, the GANs training can be done as follows. Note that, x = G(z), z p(z) and D is a set of function that holds 1-Lipschitz condition. min max E x P r [D(x)] E x Pg [D( x)] (1) G D D In order to enforce D to be a function that holds 1-Lipschitz condition, [] suggests to regularize objective function by adding a gradient penalty term. Note that Pˆx is a sampling distribution that samples from the straight line between x P r and x P g, that is, ˆx = ɛ x + (1 ɛ) x, when 0 ɛ 1 and λ g is a gradient penalty coefficient. L = E x Pg [D( x)] E x Pr [D(x)] + λ g Eˆx Pˆx [( ˆx D(ˆx) 1) ] () 31st Conference on Neural Information Processing Systems (NIPS 017), Long Beach, CA, USA.

2 3 Model setup 3.1 Objective function We define x m, x s, and x s as mixture, real source paired with mixture, and fake (generated) source paired with mixture respectively. In our setting, the goal of G is to transform x m into x s as similar as possible to x s and the goal of D is to distinguish real source x s from the fake source x s conditioned on x m. To formulate this, we changed the aforementioned objective () into conditional GANs fashion [7, 8]. Thus, the input of D becomes the concatenation of either (x m, x s ) or (x m, x s ). For the gradient penalty term, we uniformly sampled ˆx ms Pˆxms from the straight line between the concatenation of (x m, x s ) and (x m, x s ) []. L = E xm P data, x s P g [D(x m, x s )] E (xm,x s) P data [D(x m, x s )] + λ g E (xm,x s) P data, x s P g, ˆx ms Pˆxms [( ˆxms D(x m, ˆx ms ) 1) ] (3) As a final objective for generator, we added l1 loss term to check the effect of more conventional loss and experimented with three cases including the objective containing only l1 loss, only generative adversarial loss and finally a case that adds both terms together. Therefore, our final objective for each generator (L G ) and discriminator (L D ) is as follows. The coefficients for adversarial loss, gradient penalty loss and l1 loss are denoted as λ D, λ g and λ l1. L G = λ D E xm P data, x s P g [D(x m, x s )] + λ l1 E xs P r x s P g [ x s x s 1 ] () L D = λ D (E xm P data, x s P g [D(x m, x s )] E (xm,x s) P data [D(x m, x s )]) + λ g E (xm,x s) P data, x s P g, ˆx ms Pˆxms [( ˆxms D(x m, ˆx ms ) 1) ] (5) 3. Network structure for generator Our generator model is constructed as follows. First, as a deep neural network we adapt U-net structure [6]. U-net consists of encoding and decoding stages and the layers in each stage are composed of convolutional layers. In the encoding stage, inputs are encoded with convolutional layers followed by batch normalization. The inputs are encoded until it becomes a vector with a length of 08. Then, by using a fully connected layer(fc layer) we encode it to a vector with a length of 51. In the following decoding stage, the input of each layer is concatenated in channel axis by using skip connection from each corresponding layer of encoding stage. Then, the concatenated layers are decoded by deconvolutional layers followed by batch normalization. For the non-linearity functions of each convolutional and deconvolutional layer, we used leaky Relu except the last layer using Relu. The more details are described in Figure Network structure for discriminator Our discriminator model is constructed as follows. First, input either (x m, x s ) or (x m, x s ) is concatenated in channel axis. We use 5 layers of convolutional layers without batch normalization since it is invalid in gradient penalty setting []. After each batch normalization layer, we used leaky Relu as a non-linearity function except the last layer that we didn t use any non-linearity. The more details are described in Figure. One noticeable aspect of our discriminator model is the fact that we intended to make the output to have the size of [7]. This allows the each pixel value of the output to equally contribute to the Wasserstein distance we compute by simply taking the mean of the output pixel values. In this way, each pixel of the output corresponds to the each different receptive region with the same receptive size of Intuitively, we assume that this is a better idea than having a full receptive size of input size(51 18), since the receptive size is roughly a quarter of the input size, and hence, each pixel is able to make a decision over a different time-frequency region of input. Also, in practice, we found out that it is not only time consuming to train but also the train fails when the receptive size becomes bigger as the layer of discriminator becomes deeper.

3 Encode (Convolution layers) Height : 51 F 7 7 C : Width : 18 Channel: 1 C : C : Skip connections 8 reshape 1 1 F 7 7 C : 1 C : C : 18 C : 56 Decode (Deconvolution layers) Figure 1: Network structure for generator. It consists of two stages, encoding and decoding with skip connections from encoding layers. F denotes filter size, S_h denotes strides over height, S_w denotes strides over width and C denotes the output channel for the next layer. C : C : 18 C : 56 C : 1 Height : Width : 18 Channel : Figure : Network structure for discriminator. Preliminary experiments.1 Dataset The DSD100 data set was used for model training. The DSD100 consists of 50 songs as a development set and 50 songs as a test set, each consisting of mixture and four sources (vocal, bass, drums and others). All recordings are digitized with a sampling frequency of,100hz.. Mini-batch composition To train our conditional GANs model, we composed our mini-batch having two parts, one as a condition part and the other one as a target source part. It might seem natural to include only the mixtures into the condition part, but we composed the condition part to include some proportion of singing voice sources as well as the mixtures. This technique was tried due to the nature of common popular music that includes intro, interlude and outro composed only with accompaniment, and hence the term "mixture" itself does not say much about the music signal composed of both singing voice and accompaniment. Because of this reason, lots of time the target source turns out to be a zero matrix which is not good for the training of the model. Moreover, in the music of the real world, there is also a chance of singing voice appearing only by itself (e.g., Acappella). Therefore we thought 3

4 there is also a need to prepare for this situation. In most experiments, the ratio between the mixtures and the singing voices in condition part was adjusted to 7:1. This is illustrated in Figure 3. Condition Target source x " x$ " Condition Target source x " x$ " x " xx$ " " x " x " Mixture Mixture Vocal Figure 3: Composition of mini-batch used in training. Mixture, true vocal, and fake vocal from generator is denoted as x m, x s and x s, respectively..3 Pre- & post-processing As a preprocessing, the songs in the dataset are split into audio segments to have a time length of seconds with a overlap of 1 second between each audio segment. Then, we converted each stereo audio segment to a mono by taking the mean of two channels. Next, we down-sampled the audio segment to,00hz and then performed short time Fourier transform on this waveform with a window size of 10 frames and hop length of 56 frames. This setting in turn makes a segment of audio into a matrix with a size of As a post-processing, to change the final extracted vocal spectrogram as a waveform, we simply applied inverse spectrogram Fourier transform using a phase of input mixture spectrogram.. Results In Figure, we show four log magnitude spectrograms to compare the effect of generative adversarial loss and l1 loss. We found out that by using generative adversarial loss, the network tries to remove the accompaniment part more aggressively compared to the case when we only use l1 loss. Thus, we assume that one of the keys to train this model is to adjust the coefficients of l1 loss (λ l1 ) and generative adversarial loss term (λ D ). Still, we have not evaluated our algorithm with the common metrics in the music source separation task which are SDR (Source to Distortion Ratio), SIR (Source to Interference Ratio), and SAR (Source to Artifact Ratio). However, for a fair quantitative evaluation, we are planning to compare our model with the algorithm evaluation results in Signal Separation Evaluation Campaign(SiSEC) 0. The generated vocal samples of our model are available on the demo website 1. (a) (b) (c) (d) (e) Figure : Log magnitude spectrograms of (a) mixture, (b) true vocal, (c) estimated vocal using generative adversarial loss only, (d) estimated vocal using l1 loss only, and (e) estimated vocal using both generative adversarial loss and l1 loss. 1 Demo audio samples for our model are available on the website :

5 References [1] Goodfellow, I. J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A. & Bengio, Y. (01). Generative adversarial nets. Advances in neural information processing systems, [] Mao, X., Li, Q., Xie, H., Lau, R. Y., Wang, Z., & Smolley, S. P. (0). Least squares generative adversarial networks. arxiv preprint ArXiv: [3] Arjovsky, M., Chintala, S., & Bottou, L. (017). Wasserstein generative adversarial networks. In International Conference on Machine Learning, 1-3. [] Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V., & Courville, A. (017). Improved Training of Wasserstein GANs. arxiv preprint arxiv: [5] Kodali, N., Abernethy, J., Hays, J., & Kira, Z. (017). How to Train Your DRAGAN. arxiv preprint arxiv: [6] Ronneberger, O. (017). Invited Talk: U-Net Convolutional Networks for Biomedical Image Segmentation. Informatik aktuell Bildverarbeitung für die Medizin 017,3-3. [7] Isola, P., Zhu, J. Y., Zhou, T., & Efros, A. A. (0). Image-to-image translation with conditional adversarial networks.arxiv preprint arxiv: [8] Mirza, M., & Osindero, S. (01). Conditional generative adversarial nets. arxiv preprint arxiv:

GANs, GANs everywhere

GANs, GANs everywhere GANs, GANs everywhere particularly, in High Energy Physics Maxim Borisyak Yandex, NRU Higher School of Economics Generative Generative models Given samples of a random variable X find X such as: P X P

More information

Deep Generative Models. (Unsupervised Learning)

Deep Generative Models. (Unsupervised Learning) Deep Generative Models (Unsupervised Learning) CEng 783 Deep Learning Fall 2017 Emre Akbaş Reminders Next week: project progress demos in class Describe your problem/goal What you have done so far What

More information

ON ADVERSARIAL TRAINING AND LOSS FUNCTIONS FOR SPEECH ENHANCEMENT. Ashutosh Pandey 1 and Deliang Wang 1,2. {pandey.99, wang.5664,

ON ADVERSARIAL TRAINING AND LOSS FUNCTIONS FOR SPEECH ENHANCEMENT. Ashutosh Pandey 1 and Deliang Wang 1,2. {pandey.99, wang.5664, ON ADVERSARIAL TRAINING AND LOSS FUNCTIONS FOR SPEECH ENHANCEMENT Ashutosh Pandey and Deliang Wang,2 Department of Computer Science and Engineering, The Ohio State University, USA 2 Center for Cognitive

More information

arxiv: v1 [cs.lg] 20 Apr 2017

arxiv: v1 [cs.lg] 20 Apr 2017 Softmax GAN Min Lin Qihoo 360 Technology co. ltd Beijing, China, 0087 mavenlin@gmail.com arxiv:704.069v [cs.lg] 0 Apr 07 Abstract Softmax GAN is a novel variant of Generative Adversarial Network (GAN).

More information

GENERATIVE ADVERSARIAL LEARNING

GENERATIVE ADVERSARIAL LEARNING GENERATIVE ADVERSARIAL LEARNING OF MARKOV CHAINS Jiaming Song, Shengjia Zhao & Stefano Ermon Computer Science Department Stanford University {tsong,zhaosj12,ermon}@cs.stanford.edu ABSTRACT We investigate

More information

Negative Momentum for Improved Game Dynamics

Negative Momentum for Improved Game Dynamics Negative Momentum for Improved Game Dynamics Gauthier Gidel Reyhane Askari Hemmat Mohammad Pezeshki Gabriel Huang Rémi Lepriol Simon Lacoste-Julien Ioannis Mitliagkas Mila & DIRO, Université de Montréal

More information

Generative Adversarial Networks

Generative Adversarial Networks Generative Adversarial Networks SIBGRAPI 2017 Tutorial Everything you wanted to know about Deep Learning for Computer Vision but were afraid to ask Presentation content inspired by Ian Goodfellow s tutorial

More information

Lecture 14: Deep Generative Learning

Lecture 14: Deep Generative Learning Generative Modeling CSED703R: Deep Learning for Visual Recognition (2017F) Lecture 14: Deep Generative Learning Density estimation Reconstructing probability density function using samples Bohyung Han

More information

arxiv: v3 [stat.ml] 20 Feb 2018

arxiv: v3 [stat.ml] 20 Feb 2018 MANY PATHS TO EQUILIBRIUM: GANS DO NOT NEED TO DECREASE A DIVERGENCE AT EVERY STEP William Fedus 1, Mihaela Rosca 2, Balaji Lakshminarayanan 2, Andrew M. Dai 1, Shakir Mohamed 2 and Ian Goodfellow 1 1

More information

Notes on Adversarial Examples

Notes on Adversarial Examples Notes on Adversarial Examples David Meyer dmm@{1-4-5.net,uoregon.edu,...} March 14, 2017 1 Introduction The surprising discovery of adversarial examples by Szegedy et al. [6] has led to new ways of thinking

More information

Training Generative Adversarial Networks Via Turing Test

Training Generative Adversarial Networks Via Turing Test raining enerative Adversarial Networks Via uring est Jianlin Su School of Mathematics Sun Yat-sen University uangdong, China bojone@spaces.ac.cn Abstract In this article, we introduce a new mode for training

More information

Importance Reweighting Using Adversarial-Collaborative Training

Importance Reweighting Using Adversarial-Collaborative Training Importance Reweighting Using Adversarial-Collaborative Training Yifan Wu yw4@andrew.cmu.edu Tianshu Ren tren@andrew.cmu.edu Lidan Mu lmu@andrew.cmu.edu Abstract We consider the problem of reweighting a

More information

Generative Adversarial Networks, and Applications

Generative Adversarial Networks, and Applications Generative Adversarial Networks, and Applications Ali Mirzaei Nimish Srivastava Kwonjoon Lee Songting Xu CSE 252C 4/12/17 2/44 Outline: Generative Models vs Discriminative Models (Background) Generative

More information

Generative adversarial networks

Generative adversarial networks 14-1: Generative adversarial networks Prof. J.C. Kao, UCLA Generative adversarial networks Why GANs? GAN intuition GAN equilibrium GAN implementation Practical considerations Much of these notes are based

More information

Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks

Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks Emily Denton 1, Soumith Chintala 2, Arthur Szlam 2, Rob Fergus 2 1 New York University 2 Facebook AI Research Denotes equal

More information

Text2Action: Generative Adversarial Synthesis from Language to Action

Text2Action: Generative Adversarial Synthesis from Language to Action Text2Action: Generative Adversarial Synthesis from Language to Action Hyemin Ahn, Timothy Ha*, Yunho Choi*, Hwiyeon Yoo*, and Songhwai Oh Abstract In this paper, we propose a generative model which learns

More information

Some theoretical properties of GANs. Gérard Biau Toulouse, September 2018

Some theoretical properties of GANs. Gérard Biau Toulouse, September 2018 Some theoretical properties of GANs Gérard Biau Toulouse, September 2018 Coauthors Benoît Cadre (ENS Rennes) Maxime Sangnier (Sorbonne University) Ugo Tanielian (Sorbonne University & Criteo) 1 video Source:

More information

Enforcing constraints for interpolation and extrapolation in Generative Adversarial Networks

Enforcing constraints for interpolation and extrapolation in Generative Adversarial Networks Enforcing constraints for interpolation and extrapolation in Generative Adversarial Networks Panos Stinis (joint work with T. Hagge, A.M. Tartakovsky and E. Yeung) Pacific Northwest National Laboratory

More information

Generative Adversarial Networks. Presented by Yi Zhang

Generative Adversarial Networks. Presented by Yi Zhang Generative Adversarial Networks Presented by Yi Zhang Deep Generative Models N(O, I) Variational Auto-Encoders GANs Unreasonable Effectiveness of GANs GANs Discriminator tries to distinguish genuine data

More information

arxiv: v1 [eess.iv] 28 May 2018

arxiv: v1 [eess.iv] 28 May 2018 Versatile Auxiliary Regressor with Generative Adversarial network (VAR+GAN) arxiv:1805.10864v1 [eess.iv] 28 May 2018 Abstract Shabab Bazrafkan, Peter Corcoran National University of Ireland Galway Being

More information

A QUANTITATIVE MEASURE OF GENERATIVE ADVERSARIAL NETWORK DISTRIBUTIONS

A QUANTITATIVE MEASURE OF GENERATIVE ADVERSARIAL NETWORK DISTRIBUTIONS A QUANTITATIVE MEASURE OF GENERATIVE ADVERSARIAL NETWORK DISTRIBUTIONS Dan Hendrycks University of Chicago dan@ttic.edu Steven Basart University of Chicago xksteven@uchicago.edu ABSTRACT We introduce a

More information

arxiv: v3 [cs.lg] 11 Jun 2018

arxiv: v3 [cs.lg] 11 Jun 2018 Lars Mescheder 1 Andreas Geiger 1 2 Sebastian Nowozin 3 arxiv:1801.04406v3 [cs.lg] 11 Jun 2018 Abstract Recent work has shown local convergence of GAN training for absolutely continuous data and generator

More information

Open Set Learning with Counterfactual Images

Open Set Learning with Counterfactual Images Open Set Learning with Counterfactual Images Lawrence Neal, Matthew Olson, Xiaoli Fern, Weng-Keen Wong, Fuxin Li Collaborative Robotics and Intelligent Systems Institute Oregon State University Abstract.

More information

Improving Visual Semantic Embedding By Adversarial Contrastive Estimation

Improving Visual Semantic Embedding By Adversarial Contrastive Estimation Improving Visual Semantic Embedding By Adversarial Contrastive Estimation Huan Ling Department of Computer Science University of Toronto huan.ling@mail.utoronto.ca Avishek Bose Department of Electrical

More information

Which Training Methods for GANs do actually Converge?

Which Training Methods for GANs do actually Converge? Lars Mescheder 1 Andreas Geiger 1 2 Sebastian Nowozin 3 Abstract Recent work has shown local convergence of GAN training for absolutely continuous data and generator distributions. In this paper, we show

More information

Wasserstein GAN. Juho Lee. Jan 23, 2017

Wasserstein GAN. Juho Lee. Jan 23, 2017 Wasserstein GAN Juho Lee Jan 23, 2017 Wasserstein GAN (WGAN) Arxiv submission Martin Arjovsky, Soumith Chintala, and Léon Bottou A new GAN model minimizing the Earth-Mover s distance (Wasserstein-1 distance)

More information

The Success of Deep Generative Models

The Success of Deep Generative Models The Success of Deep Generative Models Jakub Tomczak AMLAB, University of Amsterdam CERN, 2018 What is AI about? What is AI about? Decision making: What is AI about? Decision making: new data High probability

More information

arxiv: v1 [cs.sd] 30 Oct 2017

arxiv: v1 [cs.sd] 30 Oct 2017 GENERATIVE ADVERSARIAL SOURCE SEPARATION Y.Cem Subaan, Paris Smaragdis, UIUC, Adobe Systems {subaan, paris}@illinois.edu arxiv:17.779v1 [cs.sd] 30 Oct 017 ABSTRACT Generative source separation methods

More information

Multiplicative Noise Channel in Generative Adversarial Networks

Multiplicative Noise Channel in Generative Adversarial Networks Multiplicative Noise Channel in Generative Adversarial Networks Xinhan Di Deepearthgo Deepearthgo@gmail.com Pengqian Yu National University of Singapore yupengqian@u.nus.edu Abstract Additive Gaussian

More information

Nishant Gurnani. GAN Reading Group. April 14th, / 107

Nishant Gurnani. GAN Reading Group. April 14th, / 107 Nishant Gurnani GAN Reading Group April 14th, 2017 1 / 107 Why are these Papers Important? 2 / 107 Why are these Papers Important? Recently a large number of GAN frameworks have been proposed - BGAN, LSGAN,

More information

ON THE DIFFERENCE BETWEEN BUILDING AND EX-

ON THE DIFFERENCE BETWEEN BUILDING AND EX- ON THE DIFFERENCE BETWEEN BUILDING AND EX- TRACTING PATTERNS: A CAUSAL ANALYSIS OF DEEP GENERATIVE MODELS. Anonymous authors Paper under double-blind review ABSTRACT Generative models are important tools

More information

Generative Adversarial Networks

Generative Adversarial Networks Generative Adversarial Networks Stefano Ermon, Aditya Grover Stanford University Lecture 10 Stefano Ermon, Aditya Grover (AI Lab) Deep Generative Models Lecture 10 1 / 17 Selected GANs https://github.com/hindupuravinash/the-gan-zoo

More information

CS 179: LECTURE 16 MODEL COMPLEXITY, REGULARIZATION, AND CONVOLUTIONAL NETS

CS 179: LECTURE 16 MODEL COMPLEXITY, REGULARIZATION, AND CONVOLUTIONAL NETS CS 179: LECTURE 16 MODEL COMPLEXITY, REGULARIZATION, AND CONVOLUTIONAL NETS LAST TIME Intro to cudnn Deep neural nets using cublas and cudnn TODAY Building a better model for image classification Overfitting

More information

Generative Adversarial Networks for Real-time Stability of Inverter-based Systems

Generative Adversarial Networks for Real-time Stability of Inverter-based Systems Generative Adversarial Networks for Real-time Stability of Inverter-based Systems Xilei Cao, Gurupraanesh Raman, Gururaghav Raman, and Jimmy Chih-Hsien Peng Department of Electrical and Computer Engineering

More information

Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net

Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net Two at Once: Enhancing Learning and Generalization Capacities via IBN-Net Supplementary Material Xingang Pan 1, Ping Luo 1, Jianping Shi 2, and Xiaoou Tang 1 1 CUHK-SenseTime Joint Lab, The Chinese University

More information

arxiv: v4 [cs.cv] 5 Sep 2018

arxiv: v4 [cs.cv] 5 Sep 2018 Wasserstein Divergence for GANs Jiqing Wu 1, Zhiwu Huang 1, Janine Thoma 1, Dinesh Acharya 1, and Luc Van Gool 1,2 arxiv:1712.01026v4 [cs.cv] 5 Sep 2018 1 Computer Vision Lab, ETH Zurich, Switzerland {jwu,zhiwu.huang,jthoma,vangool}@vision.ee.ethz.ch,

More information

arxiv: v4 [stat.ml] 16 Sep 2018

arxiv: v4 [stat.ml] 16 Sep 2018 Relaxed Wasserstein with Applications to GANs in Guo Johnny Hong Tianyi Lin Nan Yang September 9, 2018 ariv:1705.07164v4 [stat.ml] 16 Sep 2018 Abstract We propose a novel class of statistical divergences

More information

arxiv: v1 [cs.lg] 18 Dec 2017

arxiv: v1 [cs.lg] 18 Dec 2017 Squeezed Convolutional Variational AutoEncoder for Unsupervised Anomaly Detection in Edge Device Industrial Internet of Things arxiv:1712.06343v1 [cs.lg] 18 Dec 2017 Dohyung Kim, Hyochang Yang, Minki Chung,

More information

Deep Feedforward Networks

Deep Feedforward Networks Deep Feedforward Networks Liu Yang March 30, 2017 Liu Yang Short title March 30, 2017 1 / 24 Overview 1 Background A general introduction Example 2 Gradient based learning Cost functions Output Units 3

More information

Matching Adversarial Networks

Matching Adversarial Networks Matching Adversarial Networks Gellért Máttyus and Raquel Urtasun Uber Advanced Technologies Group and University of Toronto gmattyus@uber.com, urtasun@uber.com Abstract Generative Adversarial Nets (GANs)

More information

arxiv: v1 [cs.lg] 7 Nov 2017

arxiv: v1 [cs.lg] 7 Nov 2017 Theoretical Limitations of Encoder-Decoder GAN architectures Sanjeev Arora, Andrej Risteski, Yi Zhang November 8, 2017 arxiv:1711.02651v1 [cs.lg] 7 Nov 2017 Abstract Encoder-decoder GANs architectures

More information

arxiv: v3 [cs.lg] 25 Dec 2017

arxiv: v3 [cs.lg] 25 Dec 2017 Improved Training of Wasserstein GANs arxiv:1704.00028v3 [cs.lg] 25 Dec 2017 Ishaan Gulrajani 1, Faruk Ahmed 1, Martin Arjovsky 2, Vincent Dumoulin 1, Aaron Courville 1,3 1 Montreal Institute for Learning

More information

Do you like to be successful? Able to see the big picture

Do you like to be successful? Able to see the big picture Do you like to be successful? Able to see the big picture 1 Are you able to recognise a scientific GEM 2 How to recognise good work? suggestions please item#1 1st of its kind item#2 solve problem item#3

More information

Compressing deep neural networks

Compressing deep neural networks From Data to Decisions - M.Sc. Data Science Compressing deep neural networks Challenges and theoretical foundations Presenter: Simone Scardapane University of Exeter, UK Table of contents Introduction

More information

Supplementary Materials for: f-gan: Training Generative Neural Samplers using Variational Divergence Minimization

Supplementary Materials for: f-gan: Training Generative Neural Samplers using Variational Divergence Minimization Supplementary Materials for: f-gan: Training Generative Neural Samplers using Variational Divergence Minimization Sebastian Nowozin, Botond Cseke, Ryota Tomioka Machine Intelligence and Perception Group

More information

EE-559 Deep learning 10. Generative Adversarial Networks

EE-559 Deep learning 10. Generative Adversarial Networks EE-559 Deep learning 10. Generative Adversarial Networks François Fleuret https://fleuret.org/dlc/ [version of: May 17, 2018] ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE Adversarial generative models François

More information

An overview of deep learning methods for genomics

An overview of deep learning methods for genomics An overview of deep learning methods for genomics Matthew Ploenzke STAT115/215/BIO/BIST282 Harvard University April 19, 218 1 Snapshot 1. Brief introduction to convolutional neural networks What is deep

More information

arxiv: v3 [cs.lg] 10 Sep 2018

arxiv: v3 [cs.lg] 10 Sep 2018 The relativistic discriminator: a key element missing from standard GAN arxiv:1807.00734v3 [cs.lg] 10 Sep 2018 Alexia Jolicoeur-Martineau Lady Davis Institute Montreal, Canada alexia.jolicoeur-martineau@mail.mcgill.ca

More information

Predicting Deeper into the Future of Semantic Segmentation Supplementary Material

Predicting Deeper into the Future of Semantic Segmentation Supplementary Material Predicting Deeper into the Future of Semantic Segmentation Supplementary Material Pauline Luc 1,2 Natalia Neverova 1 Camille Couprie 1 Jakob Verbeek 2 Yann LeCun 1,3 1 Facebook AI Research 2 Inria Grenoble,

More information

WaveNet: A Generative Model for Raw Audio

WaveNet: A Generative Model for Raw Audio WaveNet: A Generative Model for Raw Audio Ido Guy & Daniel Brodeski Deep Learning Seminar 2017 TAU Outline Introduction WaveNet Experiments Introduction WaveNet is a deep generative model of raw audio

More information

arxiv: v2 [cs.lg] 21 Aug 2018

arxiv: v2 [cs.lg] 21 Aug 2018 CoT: Cooperative Training for Generative Modeling of Discrete Data arxiv:1804.03782v2 [cs.lg] 21 Aug 2018 Sidi Lu Shanghai Jiao Tong University steve_lu@apex.sjtu.edu.cn Weinan Zhang Shanghai Jiao Tong

More information

arxiv: v2 [cs.sd] 4 Feb 2019

arxiv: v2 [cs.sd] 4 Feb 2019 SEMI-SUPERVISED MONAURAL SINGING VOICE SEPARATION WITH A MASKING NETWORK TRAINED ON SYNTHETIC MIXTURES Michael Michelashvili 1, Sagie Benaim 1, Lior Wolf 1,2 1 Tel Aviv University 2 Facebook AI Research

More information

Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, Spis treści

Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, Spis treści Deep learning / Ian Goodfellow, Yoshua Bengio and Aaron Courville. - Cambridge, MA ; London, 2017 Spis treści Website Acknowledgments Notation xiii xv xix 1 Introduction 1 1.1 Who Should Read This Book?

More information

A Variance Modeling Framework Based on Variational Autoencoders for Speech Enhancement

A Variance Modeling Framework Based on Variational Autoencoders for Speech Enhancement A Variance Modeling Framework Based on Variational Autoencoders for Speech Enhancement Simon Leglaive 1 Laurent Girin 1,2 Radu Horaud 1 1: Inria Grenoble Rhône-Alpes 2: Univ. Grenoble Alpes, Grenoble INP,

More information

First Order Generative Adversarial Networks

First Order Generative Adversarial Networks Calvin Seward 1 2 Thomas Unterthiner 2 Urs Bergmann 1 Nikolay Jetchev 1 Sepp Hochreiter 2 Abstract GANs excel at learning high dimensional distributions, but they can update generator parameters in directions

More information

Wafer Pattern Recognition Using Tucker Decomposition

Wafer Pattern Recognition Using Tucker Decomposition Wafer Pattern Recognition Using Tucker Decomposition Ahmed Wahba, Li-C. Wang, Zheng Zhang UC Santa Barbara Nik Sumikawa NXP Semiconductors Abstract In production test data analytics, it is often that an

More information

arxiv: v1 [cs.lg] 6 Nov 2016

arxiv: v1 [cs.lg] 6 Nov 2016 GENERATIVE ADVERSARIAL NETWORKS AS VARIA- TIONAL TRAINING OF ENERGY BASED MODELS Shuangfei Zhai Binghamton University Vestal, NY 13902, USA szhai2@binghamton.edu Yu Cheng IBM T.J. Watson Research Center

More information

CAUSAL GAN: LEARNING CAUSAL IMPLICIT GENERATIVE MODELS WITH ADVERSARIAL TRAINING

CAUSAL GAN: LEARNING CAUSAL IMPLICIT GENERATIVE MODELS WITH ADVERSARIAL TRAINING CAUSAL GAN: LEARNING CAUSAL IMPLICIT GENERATIVE MODELS WITH ADVERSARIAL TRAINING (Murat Kocaoglu, Christopher Snyder, Alexandros G. Dimakis & Sriram Vishwanath, 2017) Summer Term 2018 Created for the Seminar

More information

OPTIMIZATION METHODS IN DEEP LEARNING

OPTIMIZATION METHODS IN DEEP LEARNING Tutorial outline OPTIMIZATION METHODS IN DEEP LEARNING Based on Deep Learning, chapter 8 by Ian Goodfellow, Yoshua Bengio and Aaron Courville Presented By Nadav Bhonker Optimization vs Learning Surrogate

More information

Provable Non-Convex Min-Max Optimization

Provable Non-Convex Min-Max Optimization Provable Non-Convex Min-Max Optimization Mingrui Liu, Hassan Rafique, Qihang Lin, Tianbao Yang Department of Computer Science, The University of Iowa, Iowa City, IA, 52242 Department of Mathematics, The

More information

arxiv: v1 [cs.lg] 7 Sep 2017

arxiv: v1 [cs.lg] 7 Sep 2017 Deep Learning the Physics of Transport Phenomena Amir Barati Farimani, Joseph Gomes, and Vijay S. Pande Department of Chemistry, Stanford University, Stanford, California 94305 (Dated: September 7, 2017)

More information

Experiments on the Consciousness Prior

Experiments on the Consciousness Prior Yoshua Bengio and William Fedus UNIVERSITÉ DE MONTRÉAL, MILA Abstract Experiments are proposed to explore a novel prior for representation learning, which can be combined with other priors in order to

More information

BiHMP-GAN: Bidirectional 3D Human Motion Prediction GAN

BiHMP-GAN: Bidirectional 3D Human Motion Prediction GAN BiHMP-GAN: Bidirectional 3D Human Motion Prediction GAN Jogendra Nath Kundu Maharshi Gor R. Venkatesh Babu Video Analytics Lab, Department of Computational and Data Sciences Indian Institute of Science,

More information

Understanding GANs: Back to the basics

Understanding GANs: Back to the basics Understanding GANs: Back to the basics David Tse Stanford University Princeton University May 15, 2018 Joint work with Soheil Feizi, Farzan Farnia, Tony Ginart, Changho Suh and Fei Xia. GANs at NIPS 2017

More information

Segmentation of Cell Membrane and Nucleus using Branches with Different Roles in Deep Neural Network

Segmentation of Cell Membrane and Nucleus using Branches with Different Roles in Deep Neural Network Segmentation of Cell Membrane and Nucleus using Branches with Different Roles in Deep Neural Network Tomokazu Murata 1, Kazuhiro Hotta 1, Ayako Imanishi 2, Michiyuki Matsuda 2 and Kenta Terai 2 1 Meijo

More information

Top Tagging with Lorentz Boost Networks and Simulation of Electromagnetic Showers with a Wasserstein GAN

Top Tagging with Lorentz Boost Networks and Simulation of Electromagnetic Showers with a Wasserstein GAN Top Tagging with Lorentz Boost Networks and Simulation of Electromagnetic Showers with a Wasserstein GAN Y. Rath, M. Erdmann, B. Fischer, L. Geiger, E. Geiser, J.Glombitza, D. Noll, T. Quast, M. Rieger,

More information

arxiv: v3 [stat.ml] 14 Mar 2018

arxiv: v3 [stat.ml] 14 Mar 2018 A-NICE-MC: Adversarial Training for MCMC Jiaming Song Stanford University tsong@cs.stanford.edu Shengjia Zhao Stanford University zhaosj1@cs.stanford.edu Stefano Ermon Stanford University ermon@cs.stanford.edu

More information

CONTINUOUS-TIME FLOWS FOR EFFICIENT INFER-

CONTINUOUS-TIME FLOWS FOR EFFICIENT INFER- CONTINUOUS-TIME FLOWS FOR EFFICIENT INFER- ENCE AND DENSITY ESTIMATION Anonymous authors Paper under double-blind review ABSTRACT Two fundamental problems in unsupervised learning are efficient inference

More information

Variational Autoencoders for Classical Spin Models

Variational Autoencoders for Classical Spin Models Variational Autoencoders for Classical Spin Models Benjamin Nosarzewski Stanford University bln@stanford.edu Abstract Lattice spin models are used to study the magnetic behavior of interacting electrons

More information

Neural networks and optimization

Neural networks and optimization Neural networks and optimization Nicolas Le Roux Criteo 18/05/15 Nicolas Le Roux (Criteo) Neural networks and optimization 18/05/15 1 / 85 1 Introduction 2 Deep networks 3 Optimization 4 Convolutional

More information

Variational Autoencoders (VAEs)

Variational Autoencoders (VAEs) September 26 & October 3, 2017 Section 1 Preliminaries Kullback-Leibler divergence KL divergence (continuous case) p(x) andq(x) are two density distributions. Then the KL-divergence is defined as Z KL(p

More information

arxiv: v1 [cs.lg] 6 Dec 2018

arxiv: v1 [cs.lg] 6 Dec 2018 Embedding-reparameterization procedure for manifold-valued latent variables in generative models arxiv:1812.02769v1 [cs.lg] 6 Dec 2018 Eugene Golikov Neural Networks and Deep Learning Lab Moscow Institute

More information

The Deep Ritz method: A deep learning-based numerical algorithm for solving variational problems

The Deep Ritz method: A deep learning-based numerical algorithm for solving variational problems The Deep Ritz method: A deep learning-based numerical algorithm for solving variational problems Weinan E 1 and Bing Yu 2 arxiv:1710.00211v1 [cs.lg] 30 Sep 2017 1 The Beijing Institute of Big Data Research,

More information

Understanding GANs: the LQG Setting

Understanding GANs: the LQG Setting Understanding GANs: the LQG Setting Soheil Feizi 1, Changho Suh 2, Fei Xia 1 and David Tse 1 1 Stanford University 2 Korea Advanced Institute of Science and Technology arxiv:1710.10793v1 [stat.ml] 30 Oct

More information

Improved Training of Wasserstein GANs

Improved Training of Wasserstein GANs Improved Training of Wasserstein GANs Ishaan Gulrajani 1, Faruk Ahmed 1, Martin Arjovsky 2, Vincent Dumoulin 1, Aaron Courville 1,3 1 Montreal Institute for Learning Algorithms 2 Courant Institute of Mathematical

More information

Introduction to Deep Neural Networks

Introduction to Deep Neural Networks Introduction to Deep Neural Networks Presenter: Chunyuan Li Pattern Classification and Recognition (ECE 681.01) Duke University April, 2016 Outline 1 Background and Preliminaries Why DNNs? Model: Logistic

More information

arxiv: v1 [cs.lg] 28 Dec 2017

arxiv: v1 [cs.lg] 28 Dec 2017 PixelSNAIL: An Improved Autoregressive Generative Model arxiv:1712.09763v1 [cs.lg] 28 Dec 2017 Xi Chen, Nikhil Mishra, Mostafa Rohaninejad, Pieter Abbeel Embodied Intelligence UC Berkeley, Department of

More information

Machine Learning for Computer Vision 8. Neural Networks and Deep Learning. Vladimir Golkov Technical University of Munich Computer Vision Group

Machine Learning for Computer Vision 8. Neural Networks and Deep Learning. Vladimir Golkov Technical University of Munich Computer Vision Group Machine Learning for Computer Vision 8. Neural Networks and Deep Learning Vladimir Golkov Technical University of Munich Computer Vision Group INTRODUCTION Nonlinear Coordinate Transformation http://cs.stanford.edu/people/karpathy/convnetjs/

More information

Source Separation Tutorial Mini-Series III: Extensions and Interpretations to Non-Negative Matrix Factorization

Source Separation Tutorial Mini-Series III: Extensions and Interpretations to Non-Negative Matrix Factorization Source Separation Tutorial Mini-Series III: Extensions and Interpretations to Non-Negative Matrix Factorization Nicholas Bryan Dennis Sun Center for Computer Research in Music and Acoustics, Stanford University

More information

Deep Generative Models for Graph Generation. Jian Tang HEC Montreal CIFAR AI Chair, Mila

Deep Generative Models for Graph Generation. Jian Tang HEC Montreal CIFAR AI Chair, Mila Deep Generative Models for Graph Generation Jian Tang HEC Montreal CIFAR AI Chair, Mila Email: jian.tang@hec.ca Deep Generative Models Goal: model data distribution p(x) explicitly or implicitly, where

More information

Dynamic Prediction Length for Time Series with Sequence to Sequence Networks

Dynamic Prediction Length for Time Series with Sequence to Sequence Networks Dynamic Prediction Length for Time Series with Sequence to Sequence Networks arxiv:1807.00425v1 [cs.lg] 2 Jul 2018 Mark Harmon, 1 Diego Klabjan 2 1 Department of Engineering Sciences and Applied Mathematics

More information

Composite Functional Gradient Learning of Generative Adversarial Models. Appendix

Composite Functional Gradient Learning of Generative Adversarial Models. Appendix A. Main theorem and its proof Appendix Theorem A.1 below, our main theorem, analyzes the extended KL-divergence for some β (0.5, 1] defined as follows: L β (p) := (βp (x) + (1 β)p(x)) ln βp (x) + (1 β)p(x)

More information

Convolutional Neural Networks. Srikumar Ramalingam

Convolutional Neural Networks. Srikumar Ramalingam Convolutional Neural Networks Srikumar Ramalingam Reference Many of the slides are prepared using the following resources: neuralnetworksanddeeplearning.com (mainly Chapter 6) http://cs231n.github.io/convolutional-networks/

More information

Convolutional Neural Network. Hung-yi Lee

Convolutional Neural Network. Hung-yi Lee al Neural Network Hung-yi Lee Why CNN for Image? [Zeiler, M. D., ECCV 2014] x 1 x 2 Represented as pixels x N The most basic classifiers Use 1 st layer as module to build classifiers Use 2 nd layer as

More information

Bounded Information Rate Variational Autoencoders

Bounded Information Rate Variational Autoencoders Bounded Information Rate Variational Autoencoders ABSTRACT Daniel Braithwaite Victoria University of Wellington School of Engineering and Computer Science daniel.braithwaite@ecs.vuw.ac.nz This paper introduces

More information

arxiv: v1 [cs.lg] 8 Dec 2016

arxiv: v1 [cs.lg] 8 Dec 2016 Improved generator objectives for GANs Ben Poole Stanford University poole@cs.stanford.edu Alexander A. Alemi, Jascha Sohl-Dickstein, Anelia Angelova Google Brain {alemi, jaschasd, anelia}@google.com arxiv:1612.02780v1

More information

CS 229 Project Final Report: Reinforcement Learning for Neural Network Architecture Category : Theory & Reinforcement Learning

CS 229 Project Final Report: Reinforcement Learning for Neural Network Architecture Category : Theory & Reinforcement Learning CS 229 Project Final Report: Reinforcement Learning for Neural Network Architecture Category : Theory & Reinforcement Learning Lei Lei Ruoxuan Xiong December 16, 2017 1 Introduction Deep Neural Network

More information

LEARNING TO SAMPLE WITH ADVERSARIALLY LEARNED LIKELIHOOD-RATIO

LEARNING TO SAMPLE WITH ADVERSARIALLY LEARNED LIKELIHOOD-RATIO LEARNING TO SAMPLE WITH ADVERSARIALLY LEARNED LIKELIHOOD-RATIO Chunyuan Li, Jianqiao Li, Guoyin Wang & Lawrence Carin Duke University {chunyuan.li,jianqiao.li,guoyin.wang,lcarin}@duke.edu ABSTRACT We link

More information

arxiv: v3 [stat.ml] 5 Apr 2018

arxiv: v3 [stat.ml] 5 Apr 2018 Adversarial Variational Optimization of Non-Differentiable Simulators Gilles Louppe University of Liège, Belgium Kyle Cranmer New York University, USA G.LOUPPE@ULIEGE.BE KYLE.CRANMER@NYU.EDU arxiv:1707.07113v3

More information

TUTORIAL PART 1 Unsupervised Learning

TUTORIAL PART 1 Unsupervised Learning TUTORIAL PART 1 Unsupervised Learning Marc'Aurelio Ranzato Department of Computer Science Univ. of Toronto ranzato@cs.toronto.edu Co-organizers: Honglak Lee, Yoshua Bengio, Geoff Hinton, Yann LeCun, Andrew

More information

From perceptrons to word embeddings. Simon Šuster University of Groningen

From perceptrons to word embeddings. Simon Šuster University of Groningen From perceptrons to word embeddings Simon Šuster University of Groningen Outline A basic computational unit Weighting some input to produce an output: classification Perceptron Classify tweets Written

More information

When Harmonic Analysis Meets Machine Learning: Lipschitz Analysis of Deep Convolution Networks

When Harmonic Analysis Meets Machine Learning: Lipschitz Analysis of Deep Convolution Networks When Harmonic Analysis Meets Machine Learning: Lipschitz Analysis of Deep Convolution Networks Radu Balan Department of Mathematics, AMSC, CSCAMM and NWC University of Maryland, College Park, MD Joint

More information

Stochastic Video Prediction with Deep Conditional Generative Models

Stochastic Video Prediction with Deep Conditional Generative Models Stochastic Video Prediction with Deep Conditional Generative Models Rui Shu Stanford University ruishu@stanford.edu Abstract Frame-to-frame stochasticity remains a big challenge for video prediction. The

More information

Theories of Deep Learning

Theories of Deep Learning Theories of Deep Learning Lecture 02 Donoho, Monajemi, Papyan Department of Statistics Stanford Oct. 4, 2017 1 / 50 Stats 385 Fall 2017 2 / 50 Stats 285 Fall 2017 3 / 50 Course info Wed 3:00-4:20 PM in

More information

arxiv: v2 [stat.ml] 23 Mar 2018 Abstract

arxiv: v2 [stat.ml] 23 Mar 2018 Abstract Published as a conference paper at CVPR 2018 Structured Uncertainty Prediction Networks Garoe Dorta 1,2 Sara Vicente 2 Lourdes Agapito 3 Neill D.F. Campbell 1 Ivor Simpson 2 1 University of Bath 2 Anthropics

More information

CSC321 Lecture 20: Reversible and Autoregressive Models

CSC321 Lecture 20: Reversible and Autoregressive Models CSC321 Lecture 20: Reversible and Autoregressive Models Roger Grosse Roger Grosse CSC321 Lecture 20: Reversible and Autoregressive Models 1 / 23 Overview Four modern approaches to generative modeling:

More information

Generating Text via Adversarial Training

Generating Text via Adversarial Training Generating Text via Adversarial Training Yizhe Zhang, Zhe Gan, Lawrence Carin Department of Electronical and Computer Engineering Duke University, Durham, NC 27708 {yizhe.zhang,zhe.gan,lcarin}@duke.edu

More information

Domain adaptation for deep learning

Domain adaptation for deep learning What you saw is not what you get Domain adaptation for deep learning Kate Saenko Successes of Deep Learning in AI A Learning Advance in Artificial Intelligence Rivals Human Abilities Deep Learning for

More information

Adversarial Examples Generation and Defense Based on Generative Adversarial Network

Adversarial Examples Generation and Defense Based on Generative Adversarial Network Adversarial Examples Generation and Defense Based on Generative Adversarial Network Fei Xia (06082760), Ruishan Liu (06119690) December 15, 2016 1 Abstract We propose a novel generative adversarial network

More information

Statistical Machine Learning

Statistical Machine Learning Statistical Machine Learning Lecture 9 Numerical optimization and deep learning Niklas Wahlström Division of Systems and Control Department of Information Technology Uppsala University niklas.wahlstrom@it.uu.se

More information