We consider the problem of minimizing a functional over a parametric family of probability measures, where the parameterization is characterized via a push-forward structure. An important application of this problem is in training generative adversarial networks. In this regard, we propose a novel Sinkhorn Natural Gradient (SiNG) algorithm which acts as a steepest descent method on the probability space endowed with the Sinkhorn divergence. We show that the Sinkhorn information matrix (SIM), a key component of SiNG, has an explicit expression and can be evaluated accurately in complexity that scales logarithmically with respect to the desired accuracy. This is in sharp contrast to existing natural gradient methods that can only be carried out approximately. Moreover, in practical applications when only Monte-Carlo type integration is available, we design an empirical estimator for SIM and provide the stability analysis. In our experiments, we quantitatively compare SiNG with state-of-the-art SGD-type solvers on generative tasks to demonstrate its efficiency and efficacy of our method.
Warsaw University of Technology（华沙理工大学）；CERN（欧洲核子研究组织）；Tooploox
In this work, we propose a novel end-to-end sinkhorn autoencoder with noise generator for efficient data collection simulation. Simulating processes that aim at collecting experimental data is crucial for multiple real-life applications, including nuclear medicine, astronomy and high energy physics. Contemporary methods, such as Monte Carlo algorithms, provide high-fidelity results at a price of high computational cost. Multiple attempts are taken to reduce this burden, e.g. using generative approaches based on Generative Adversarial Networks or Variational Autoencoders. Although such methods …
In this paper, we consider the problem of computing the barycenter of a set of probability distributions under the Sinkhorn divergence. This problem has recently found applications across various domains, including graphics, learning, and vision, as it provides a meaningful mechanism to aggregate knowledge. Unlike previous approaches which directly operate in the space of probability measures, we recast the Sinkhorn barycenter problem as an instance of unconstrained functional optimization and develop a novel functional gradient descent method named Sinkhorn Descent (SD). We prove that SD conv…
Sinkhorn Hamiltonian Monte Carlo for Entropic Optimal Transport Generalized Bayes
University of Technology of Troyes（特鲁瓦技术大学）；QAD Inc.（QAD公司）；Conservatoire national des arts et métiers（国立工艺院）；CNRS 6051, RSMS, EHESP（法国国家科学研究中心6051、公共卫生研究与管理学院、高等卫生与医学教育研究所）
Many problems at the intersection of combinatorics and computer science require solving for a permutation that optimally matches, ranks, or sorts some data. These problems usually have a task-specific, often non-differentiable objective function that data-driven algorithms can use as a learning signal. In this paper, we propose the Sinkhorn Policy Gradient (SPG) algorithm for learning policies on permutation matrices. The actor-critic neural network architecture we introduce for SPG uniquely decouples representation learning of the state space from the highly-structured action space of permuta…
RJ Research Consulting（RJ研究咨询公司）；Tencent AI Lab（腾讯人工智能实验室）
This paper first presents a theory for generative adversarial methods that does not rely on the traditional minimax formulation. It shows that with a strong discriminator, a good generator can be learned so that the KL divergence between the distributions of real data and generated data improves after each functional gradient step until it converges to zero. Based on the theory, we propose a new stable generative adversarial method. A theoretical insight into the original GAN from this new viewpoint is also provided. The experiments on image generation show the effectiveness of our new method.
Oregon State University（俄勒冈州立大学）；Horizon Robotics（Horizon Robotics（地平线机器人））
Recently, particle-based variational inference (ParVI) methods have gained interest because they can avoid arbitrary parametric assumptions that are common in variational inference. However, many ParVI approaches do not allow arbitrary sampling from the posterior, and the few that do allow such sampling suffer from suboptimality. This work proposes a new method for learning to approximately sample from the posterior distribution. We construct a neural sampler that is trained with the functional gradient of the KL-divergence between the empirical sampling distribution and the target distributio…
Generative Neural Networks for Sinkhorn Distributionally Robust Hypothesis Testing
School of Artificial Intelligence, The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳）人工智能学院）；College of Business, City University of Hong Kong（香港城市大学商学院）；School of Data Science, The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳）数据科学学院）
We introduce a new generative model where samples are produced via Langevin dynamics using gradients of the data distribution estimated with score matching. Because gradients can be ill-defined and hard to estimate when the data resides on low-dimensional manifolds, we perturb the data with different levels of Gaussian noise, and jointly estimate the corresponding scores, i.e., the vector fields of gradients of the perturbed data distribution for all noise levels. For sampling, we propose an annealed Langevin dynamics where we use gradients corresponding to gradually decreasing noise levels as…
Guiding Time-Varying Generative Models with Natural Gradients on Exponential Family Manifold
Song Liu School of Mathematics University of Bristol United Kingdom（索恩·刘 布里斯托大学数学学院）；Leyang Wang Department of Computer Science University College London United Kingdom（雷杨·王 伦敦大学学院计算机科学系）；Yakun Wang School of Mathematics University of Bristol United Kingdom（亚坤·王 布里斯托大学数学学院）
University College London（伦敦大学学院）；INRIA（法国国家信息与自动化研究所）；École Normale Supérieure（巴黎高等师范学院）；Istituto Italiano di Tecnologia（意大利技术研究院）
Applications of optimal transport have recently gained remarkable attention thanks to the computational advantages of entropic regularization. However, in most situations the Sinkhorn approximation of the Wasserstein distance is replaced by a regularized version that is less accurate but easy to differentiate. In this work we characterize the differential properties of the original Sinkhorn distance, proving that it enjoys the same smoothness as its regularized version and we explicitly provide an efficient algorithm to compute its gradient. We show that this result benefits both theory and ap…
The Hong Kong University of Science and Technology（香港科技大学）
This paper considers the general $f$-divergence formulation of bidirectional generative modeling, which includes VAE and BiGAN as special cases. We present a new optimization method for this formulation, where the gradient is computed using an adversarially learned discriminator. In our framework, we show that different divergences induce similar algorithms in terms of gradient evaluation, except with different scaling. Therefore this paper gives a general recipe for a class of principled $f$-divergence based generative modeling methods. Theoretical justifications and extensive empirical studi…
Rutgers University（罗格斯大学）；University of Patras（帕特雷大学）
Generative adversarial networks (GANs) are designed with the help of min-max optimization problems that are solved with stochastic gradient-type algorithms which are known to be non-robust. In this work we revisit a non-adversarial method based on kernels which relies on a pure minimization problem and propose a simple stochastic gradient algorithm for the computation of its solution. Using simplified tools from Stochastic Approximation theory we demonstrate that batch versions of the algorithm or smoothing of the gradient do not improve convergence. These observations allow for the developmen…
The ability to compare two degenerate probability distributions (i.e. two probability distributions supported on two distinct low-dimensional manifolds living in a much higher-dimensional space) is a crucial problem arising in the estimation of generative models for high-dimensional observations such as those arising in computer vision or natural language. It is known that optimal transport metrics can represent a cure for this problem, since they were specifically designed as an alternative to information divergences to handle such problematic scenarios. Unfortunately, training generative mac…
Xi’an Jiaotong University（西安交通大学）；Zhongnan University of Economics and Law（中南财经政法大学）；The Hong Kong University of Science and Technology（香港科技大学）
We propose a general framework to learn deep generative models via \textbf{V}ariational \textbf{Gr}adient Fl\textbf{ow} (VGrow) on probability spaces. The evolving distribution that asymptotically converges to the target distribution is governed by a vector field, which is the negative gradient of the first variation of the $f$-divergence between them. We prove that the evolving distribution coincides with the pushforward distribution through the infinitesimal time composition of residual maps that are perturbations of the identity map along the vector field. The vector field depends on the de…
In this paper, we propose a novel normalization method called gradient normalization (GN) to tackle the training instability of Generative Adversarial Networks (GANs) caused by the sharp gradient space. Unlike existing work such as gradient penalty and spectral normalization, the proposed GN only imposes a hard 1-Lipschitz constraint on the discriminator function, which increases the capacity of the discriminator. Moreover, the proposed gradient normalization can be applied to different GAN architectures with little modification. Extensive experiments on four datasets show that GANs trained wi…
Deep generative modeling has seen impressive advances in recent years, to the point where it is now commonplace to see simulated samples (e.g., images) that closely resemble real-world data. However, generation quality is generally inconsistent for any given model and can vary dramatically between samples. We introduce Discriminator Gradient flow (DGflow), a new technique that improves generated samples via the gradient flow of entropy-regularized f-divergences between the real and the generated data distributions. The gradient flow takes the form of a non-linear Fokker-Plank equation, which c…
While Generative Adversarial Networks (GANs) have seen huge successes in image synthesis tasks, they are notoriously difficult to adapt to different datasets, in part due to instability during training and sensitivity to hyperparameters. One commonly accepted reason for this instability is that gradients passing from the discriminator to the generator become uninformative when there isn't enough overlap in the supports of the real and fake distributions. In this work, we propose the Multi-Scale Gradient Generative Adversarial Network (MSG-GAN), a simple but effective technique for addressing t…
King Abdullah University of Science and Technology（阿卜杜拉国王科技大学）
We present a new perspective on the celebrated Sinkhorn algorithm by showing that is a special case of incremental/stochastic mirror descent. In order to see this, one should simply plug Kullback-Leibler divergence in both mirror map and the objective function. Since the problem has unbounded domain, the objective function is neither smooth nor it has bounded gradients. However, one can still approach the problem using the notion of relative smoothness, obtaining that the stochastic objective is 1-relative smooth. The discovered equivalence allows us to propose 1) new methods for optimal trans…
University of Amsterdam（阿姆斯特丹大学）；Deeptrace；Google Brain（谷歌大脑）；DeepMind（深度思考）；École Polytechnique（巴黎综合理工学院）；Sony CSL（索尼计算机科学实验室）
Optimal transport offers an alternative to maximum likelihood for learning generative autoencoding models. We show that minimizing the p-Wasserstein distance between the generator and the true data distribution is equivalent to the unconstrained min-min optimization of the p-Wasserstein distance between the encoder aggregated posterior and the prior in latent space, plus a reconstruction error. We also identify the role of its trade-off hyperparameter as the capacity of the generator: its Lipschitz constant. Moreover, we prove that optimizing the encoder over any class of universal approximato…
Intractable generative models are models for which the likelihood is unavailable but sampling is possible. Most approaches to parameter inference in this setting require the computation of some discrepancy between the data and the generative model. This is for example the case for minimum distance estimation and approximate Bayesian computation. These approaches require sampling a high number of realisations from the model for different parameter values, which can be a significant challenge when simulating is an expensive operation. In this paper, we propose to enhance this approach by enforci…
Cohomological Obstructions to Global Counterfactuals: A Sheaf-Theoretic Foundation for Generative Causal Models
School of Management, University of Science and Technology of China（管理学院，中国科学技术大学）；School of Computer Science and Engineering, University of Science and Technology of China（计算机科学与工程学院，中国科学技术大学）
Bioinformatics Institute, A*STAR（新加坡科技研究局生物信息学研究所）；Huaiyin Normal University（淮阴师范学院）；Nanjing Normal University（南京师范大学）；University of Alberta（阿尔伯塔大学）
Generative adversarial networks (GANs), famous for the capability of learning complex underlying data distribution, are however known to be tricky in the training process, which would probably result in mode collapse or performance deterioration. Current approaches of dealing with GANs' issues almost utilize some practical training techniques for the purpose of regularization, which on the other hand undermines the convergence and theoretical soundness of GAN. In this paper, we propose to stabilize GAN training via a novel particle-based variational inference -- Langevin Stein variational grad…
Generative Model Inversion Through the Lens of the Manifold Hypothesis
TMLR Group, Department of Computer Science, Hong Kong Baptist University（香港 Baptist 大学计算机科学系）；Sydney AI Centre, The University of Sydney（悉尼大学人工智能中心）；School of Computing and Information Systems, The University of Melbourne（墨尔本大学计算与信息系）；McCombs School of Business, The University of Texas at Austin（德克萨斯大学奥斯汀分校麦科姆斯商学院）
Deep generative models (e.g. GANs and VAEs) have been developed quite extensively in recent years. Lately, there has been an increased interest in the inversion of such a model, i.e. given a (possibly corrupted) signal, we wish to recover the latent vector that generated it. Building upon sparse representation theory, we define conditions that are applicable to any inversion algorithm (gradient descent, deep encoder, etc.), under which such generative models are invertible with a unique solution. Importantly, the proposed analysis is applicable to any trained model, and does not depend on Gaus…
We are interested in gradient-based Explicit Generative Modeling where samples can be derived from iterative gradient updates based on an estimate of the score function of the data distribution. Recent advances in Stochastic Gradient Langevin Dynamics (SGLD) demonstrates impressive results with energy-based models on high-dimensional and complex data distributions. Stein Variational Gradient Descent (SVGD) is a deterministic sampling algorithm that iteratively transports a set of particles to approximate a given distribution, based on functional gradient descent that decreases the KL divergenc…
In this paper, we study the convergence of generative adversarial networks (GANs) from the perspective of the informativeness of the gradient of the optimal discriminative function. We show that GANs without restriction on the discriminative function space commonly suffer from the problem that the gradient produced by the discriminator is uninformative to guide the generator. By contrast, Wasserstein GAN (WGAN), where the discriminative function is restricted to 1-Lipschitz, does not suffer from such a gradient uninformativeness problem. We further show in the paper that the model with a compa…
Mila（米拉人工智能研究所）；Université de Montréal（蒙特利尔大学）；Meiji University（明治大学）
One of the training strategies of generative models is to minimize the Jensen--Shannon divergence between the model distribution and the data distribution. Since data distribution is unknown, generative adversarial networks (GANs) formulate this problem as a game between two models, a generator and a discriminator. The training can be formulated in the context of game theory and the local Nash equilibrium (LNE). It does not seem feasible to derive guarantees of stability or optimality for the existing methods. This optimization problem is far more challenging than the single objective setting.…
Shenzhen University（深圳大学）；Tencent AI Lab（腾讯人工智能实验室）
Training generative adversarial networks (GAN) in a distributed fashion is a promising technology since it is contributed to training GAN on a massive of data efficiently in real-world applications. However, GAN is known to be difficult to train by SGD-type methods (may fail to converge) and the distributed SGD-type methods may also suffer from massive amount of communication cost. In this paper, we propose a {distributed GANs training algorithm with quantized gradient, dubbed DQGAN,} which is the first distributed training method with quantized gradient for GANs. The new method trains GANs ba…
Department of Applied Mathematics and Theoretical Physics, University of Cambridge（应用数学与理论物理系，剑桥大学）；Nonlinear Systems Laboratory, Massachusetts Institute of Technology（非线性系统实验室，麻省理工学院）
CEREMADE, Université Paris Dauphine - PSL（巴黎第九大学CEREMADE）
Fitting neural networks often resorts to stochastic (or similar) gradient descent which is a noise-tolerant (and efficient) resolution of a gradient descent dynamics. It outputs a sequence of networks parameters, which sequence evolves during the training steps. The gradient descent is the limit, when the learning rate is small and the batch size is infinite, of this set of increasingly optimal network parameters obtained during training. In this contribution, we investigate instead the convergence in the Generative Adversarial Networks used in machine learning. We study the limit of small lea…
The University of Iowa（爱荷华大学）；IBM T. J. Watson Research Center（IBM托马斯·J·沃森研究中心）
Adaptive gradient algorithms perform gradient-based updates using the history of gradients and are ubiquitous in training deep neural networks. While adaptive gradient methods theory is well understood for minimization problems, the underlying factors driving their empirical success in min-max problems such as GANs remain unclear. In this paper, we aim at bridging this gap from both theoretical and empirical perspectives. First, we analyze a variant of Optimistic Stochastic Gradient (OSG) proposed in~\citep{daskalakis2017training} for solving a class of non-convex non-concave min-max problem a…
Quantum machine learning is expected to be one of the first potential general-purpose applications of near-term quantum devices. A major recent breakthrough in classical machine learning is the notion of generative adversarial training, where the gradients of a discriminator model are used to train a separate generative model. In this work and a companion paper, we extend adversarial training to the quantum domain and show how to construct generative adversarial networks using quantum circuits. Furthermore, we also show how to compute gradients -- a key element in generative adversarial networ…
Gaussian entropic optimal transport: Schrödinger bridges and the Sinkhorn algorithm
Department of Mathematics, Imperial College London, UK；Centre de Recherche Inria Bordeaux Sud-Ouest, Talence, France；Department of Signal Theory \& Communications, Universidad Carlos III de Madrid, Spain
The wide applications of Generative adversarial networks benefit from the successful training methods, guaranteeing that an object function converges to the local minima. Nevertheless, designing an efficient and competitive training method is still a challenging task due to the cyclic behaviors of some gradient-based ways and the expensive computational cost of these methods based on the Hessian matrix. This paper proposed the adaptive Composite Gradients (ACG) method, linearly convergent in bilinear games under suitable settings. Theory and toy-function experiments suggest that our approach c…
Iowa State University（爱荷华州立大学）；New York University（纽约大学）
Deep generative models have emerged as a powerful class of priors for signals in various inverse problems such as compressed sensing, phase retrieval and super-resolution. Here, we assume an unknown signal to lie in the range of some pre-trained generative model. A popular approach for signal recovery is via gradient descent in the low-dimensional latent space. While gradient descent has achieved good empirical performance, its theoretical behavior is not well understood. In this paper, we introduce the use of stochastic gradient Langevin dynamics (SGLD) for compressed sensing with a generativ…
Generative adversarial imitation learning (GAIL) is a popular inverse reinforcement learning approach for jointly optimizing policy and reward from expert trajectories. A primary question about GAIL is whether applying a certain policy gradient algorithm to GAIL attains a global minimizer (i.e., yields the expert policy), for which existing understanding is very limited. Such global convergence has been shown only for the linear (or linear-type) MDP and linear (or linearizable) reward. In this paper, we study GAIL under general MDP and for nonlinear reward function classes (as long as the obje…
We propose a new technique that boosts the convergence of training generative adversarial networks. Generally, the rate of training deep models reduces severely after multiple iterations. A key reason for this phenomenon is that a deep network is expressed using a highly non-convex finite-dimensional model, and thus the parameter gets stuck in a local optimum. Because of this, methods often suffer not only from degeneration of the convergence speed but also from limitations in the representational power of the trained network. To overcome this issue, we propose an additional layer called the g…
