arXiv subjects

Matthew D. Hoffman

Publications and source records attributed to Matthew D. Hoffman.

30 records · Page 2Linked to original sources

arXiv2018

Stochastic Gradient Descent as Approximate Bayesian Inference

Stochastic Gradient Descent with a constant learning rate (constant SGD) simulates a Markov chain with a stationary distribution. With this perspective, we derive several new results. (1) We show that constant SGD can be used as an approximate Bayesian posterior inference algorithm. Specifically, we show how to adjust the tuning parameters of constant SGD to best match the stationary distribution to a posterior, minimizing the Kullback-Leibler divergence between these two distributions. (2) We demonstrate that constant SGD gives rise to a new variational EM algorithm that optimizes hyperparameters in complex probabilistic models. (3) We also propose SGD with momentum for sampling and show how to adjust the damping coefficient accordingly. (4) We analyze MCMC algorithms. For Langevin Dynamics and Stochastic Gradient Fisher Scoring, we quantify the approximation errors due to finite learning rates. Finally (5), we use the stochastic process perspective to give a short proof of why Polyak averaging is optimal. Based on this idea, we propose a scalable approximate MCMC algorithm, the Averaged Stochastic Gradient Sampler.

stat.ML↗

arXiv2017

Multimodal Prediction and Personalization of Photo Edits with Deep Generative Models

Professional-grade software applications are powerful but complicated$-$expert users can achieve impressive results, but novices often struggle to complete even basic tasks. Photo editing is a prime example: after loading a photo, the user is confronted with an array of cryptic sliders like "clarity", "temp", and "highlights". An automatically generated suggestion could help, but there is no single "correct" edit for a given image$-$different experts may make very different aesthetic decisions when faced with the same image, and a single expert may make different choices depending on the intended use of the image (or on a whim). We therefore want a system that can propose multiple diverse, high-quality edits while also learning from and adapting to a user's aesthetic preferences. In this work, we develop a statistical model that meets these objectives. Our model builds on recent advances in neural network generative modeling and scalable inference, and uses hierarchical structure to learn editing patterns across many diverse users. Empirically, we find that our model outperforms other approaches on this challenging multimodal prediction task.

stat.ML↗

arXiv2017

Deep Probabilistic Programming

We propose Edward, a Turing-complete probabilistic programming language. Edward defines two compositional representations---random variables and inference. By treating inference as a first class citizen, on a par with modeling, we show that probabilistic programming can be as flexible and computationally efficient as traditional deep learning. For flexibility, Edward makes it easy to fit the same model using a variety of composable inference methods, ranging from point estimation to variational inference to MCMC. In addition, Edward can reuse the modeling representation as part of inference, facilitating the design of rich variational models and generative adversarial networks. For efficiency, Edward is integrated into TensorFlow, providing significant speedups over existing probabilistic systems. For example, we show on a benchmark logistic regression task that Edward is at least 35x faster than Stan and 6x faster than PyMC3. Further, Edward incurs no runtime overhead: it is as fast as handwritten TensorFlow.

stat.ML↗

arXiv2016

A Variational Analysis of Stochastic Gradient Algorithms

Stochastic Gradient Descent (SGD) is an important algorithm in machine learning. With constant learning rates, it is a stochastic process that, after an initial phase of convergence, generates samples from a stationary distribution. We show that SGD with constant rates can be effectively used as an approximate posterior inference algorithm for probabilistic modeling. Specifically, we show how to adjust the tuning parameters of SGD such as to match the resulting stationary distribution to the posterior. This analysis rests on interpreting SGD as a continuous-time stochastic process and then minimizing the Kullback-Leibler divergence between its stationary distribution and the target posterior. (This is in the spirit of variational inference.) In more detail, we model SGD as a multivariate Ornstein-Uhlenbeck process and then use properties of this process to derive the optimal parameters. This theoretical framework also connects SGD to modern scalable inference algorithms; we analyze the recently proposed stochastic gradient Fisher scoring under this perspective. We demonstrate that SGD with properly chosen constant rates gives a new way to optimize hyperparameters in probabilistic models.

stat.ML↗

arXiv2015

The Stan Math Library: Reverse-Mode Automatic Differentiation in C++

As computational challenges in optimization and statistical inference grow ever harder, algorithms that utilize derivatives are becoming increasingly more important. The implementation of the derivatives that make these algorithms so powerful, however, is a substantial user burden and the practicality of these algorithms depends critically on tools like automatic differentiation that remove the implementation burden entirely. The Stan Math Library is a C++, reverse-mode automatic differentiation library designed to be usable, extensive and extensible, efficient, scalable, stable, portable, and redistributable in order to facilitate the construction and utilization of such algorithms. Usability is achieved through a simple direct interface and a cleanly abstracted functional interface. The extensive built-in library includes functions for matrix operations, linear algebra, differential equation solving, and most common probability functions. Extensibility derives from a straightforward object-oriented framework for expressions, allowing users to easily create custom functions. Efficiency is achieved through a combination of custom memory management, subexpression caching, traits-based metaprogramming, and expression templates. Partial derivatives for compound functions are evaluated lazily for improved scalability. Stability is achieved by taking care with arithmetic precision in algebraic expressions and providing stable, compound functions where possible. For portability, the library is standards-compliant C++ (03) and has been tested for all major compilers for Windows, Mac OS X, and Linux.

cs.MS↗

arXiv2015

A trust-region method for stochastic variational inference with applications to streaming data

Stochastic variational inference allows for fast posterior inference in complex Bayesian models. However, the algorithm is prone to local optima which can make the quality of the posterior approximation sensitive to the choice of hyperparameters and initialization. We address this problem by replacing the natural gradient step of stochastic varitional inference with a trust-region update. We show that this leads to generally better results and reduced sensitivity to hyperparameters. We also describe a new strategy for variational inference on streaming data and show that here our trust-region method is crucial for getting good performance.

stat.ML↗

arXiv2014

Beta Process Non-negative Matrix Factorization with Stochastic Structured Mean-Field Variational Inference

Beta process is the standard nonparametric Bayesian prior for latent factor model. In this paper, we derive a structured mean-field variational inference algorithm for a beta process non-negative matrix factorization (NMF) model with Poisson likelihood. Unlike the linear Gaussian model, which is well-studied in the nonparametric Bayesian literature, NMF model with beta process prior does not enjoy the conjugacy. We leverage the recently developed stochastic structured mean-field variational inference to relax the conjugacy constraint and restore the dependencies among the latent variables in the approximating variational distribution. Preliminary results on both synthetic and real examples demonstrate that the proposed inference algorithm can reasonably recover the hidden structure of the data.

stat.ML↗

arXiv2014

Structured Stochastic Variational Inference

Stochastic variational inference makes it possible to approximate posterior distributions induced by large datasets quickly using stochastic optimization. The algorithm relies on the use of fully factorized variational distributions. However, this "mean-field" independence approximation limits the fidelity of the posterior approximation, and introduces local optima. We show how to relax the mean-field approximation to allow arbitrary dependencies between global parameters and local hidden variables, producing better parameter estimates by reducing bias, sensitivity to local optima, and sensitivity to hyperparameters.

cs.LG↗

arXiv2014

A Generative Product-of-Filters Model of Audio

We propose the product-of-filters (PoF) model, a generative model that decomposes audio spectra as sparse linear combinations of "filters" in the log-spectral domain. PoF makes similar assumptions to those used in the classic homomorphic filtering approach to signal processing, but replaces hand-designed decompositions built of basic signal processing operations with a learned decomposition based on statistical inference. This paper formulates the PoF model and derives a mean-field method for posterior inference and a variational EM algorithm to estimate the model's free parameters. We demonstrate PoF's potential for audio processing on a bandwidth expansion task, and show that PoF can serve as an effective unsupervised feature extractor for a speaker identification task.

stat.ML↗

arXiv2014

Image Classification and Retrieval from User-Supplied Tags

This paper proposes direct learning of image classification from user-supplied tags, without filtering. Each tag is supplied by the user who shared the image online. Enormous numbers of these tags are freely available online, and they give insight about the image categories important to users and to image classification. Our approach is complementary to the conventional approach of manual annotation, which is extremely costly. We analyze of the Flickr 100 Million Image dataset, making several useful observations about the statistics of these tags. We introduce a large-scale robust classification algorithm, in order to handle the inherent noise in these tags, and a calibration procedure to better predict objective annotations. We show that freely available, user-supplied tags can obtain similar or superior results to large databases of costly manual annotations.

cs.CV↗

arXiv2011

The No-U-Turn Sampler: Adaptively Setting Path Lengths in Hamiltonian Monte Carlo

Hamiltonian Monte Carlo (HMC) is a Markov chain Monte Carlo (MCMC) algorithm that avoids the random walk behavior and sensitivity to correlated parameters that plague many MCMC methods by taking a series of steps informed by first-order gradient information. These features allow it to converge to high-dimensional target distributions much more quickly than simpler methods such as random walk Metropolis or Gibbs sampling. However, HMC's performance is highly sensitive to two user-specified parameters: a step size ε and a desired number of steps L. In particular, if L is too small then the algorithm exhibits undesirable random walk behavior, while if L is too large the algorithm wastes computation. We introduce the No-U-Turn Sampler (NUTS), an extension to HMC that eliminates the need to set a number of steps L. NUTS uses a recursive algorithm to build a set of likely candidate points that spans a wide swath of the target distribution, stopping automatically when it starts to double back and retrace its steps. Empirically, NUTS perform at least as efficiently as and sometimes more efficiently than a well tuned standard HMC method, without requiring user intervention or costly tuning runs. We also derive a method for adapting the step size parameter ε on the fly based on primal-dual averaging. NUTS can thus be used with no hand-tuning at all. NUTS is also suitable for applications such as BUGS-style automatic inference engines that require efficient "turnkey" sampling algorithms.

stat.CO↗

arXiv2010

Approximate Maximum A Posteriori Inference with Entropic Priors

In certain applications it is useful to fit multinomial distributions to observed data with a penalty term that encourages sparsity. For example, in probabilistic latent audio source decomposition one may wish to encode the assumption that only a few latent sources are active at any given time. The standard heuristic of applying an L1 penalty is not an option when fitting the parameters to a multinomial distribution, which are constrained to sum to 1. An alternative is to use a penalty term that encourages low-entropy solutions, which corresponds to maximum a posteriori (MAP) parameter estimation with an entropic prior. The lack of conjugacy between the entropic prior and the multinomial distribution complicates this approach. In this report I propose a simple iterative algorithm for MAP estimation of multinomial distributions with sparsity-inducing entropic priors.

cs.SD↗