Score function estimators improve k-subset sampling efficiency.
problem Efficiently sampling k-subsets in machine learning tasks. method Revisit score function estimators, using discrete Fourier transform and control variates.
result Efficient and unbiased gradient estimates for k-subset sampling. Paper presents a new policy gradient theorem using weak derivatives for reinforcement learning.
problem Continuous state-action reinforcement learning problems.
method Introduced an alternative policy gradient theorem using weak derivatives.
result The new approach yields algorithms that converge almost surely to stationary points of the value function.
Paper presents variational estimates for EBLVMs without structural assumptions.
problem Challenges in learning and evaluating EBLVMs due to intractable true posteriors and partition functions.
method Variational estimates of the score function and its gradient (VaES and VaGES) in a general EBLVM.
result The estimates can be applied to KSD and SM-based methods to learn EBLVMs and estimate Fisher divergence.
New maximum score estimators using ReLU functions and deep neural networks.
problem Estimating parameters in models with sign restrictions.
method ReLU-based maximum score criterion and DNN architecture.
result RMS estimator achieves n−s/(2s+1) convergence rate and asymptotic normality. The paper establishes bounds for score-matching in causal discovery and generative modeling.
problem Estimating causal relationships from data.
method Training a deep neural network to estimate the score function and applying it to causal discovery.
result Bounds on the error rate of causal discovery methods using score-matching.
A new method improves stochastic gradient descent for faster and more efficient estimation.
problem Efficient and fast parametric estimation methods.
method Projected stochastic gradient descent corrected by Fisher scoring.
result The method is faster and more efficient than traditional methods.
Paper improves variance control in importance weighted variational bounds.
problem Improving the variance of gradient estimators for IWAE.
method Develops a novel control variate that grows SNR as √K for large K.
result Empirically, the method yields superior variance reduction for generative models.
A new method for discrete data normalizing flows using latent transformations.
problem Challenges in parameterizing bijective transformations for discrete data.
method Predict a distribution over latent transformations to make the marginal likelihood differentiable.
result Discrete-data normalizing flows can be trained using gradient-based learning with unbiased score function estimation.
This work proposes using zero-variance control variates to reduce variance in pathwise gradient estimators for variational inference.
problem Pathwise gradient estimators in variational inference have high variance, leading to inefficient optimization.
method Apply zero-variance control variates to pathwise gradient estimators.
result Zero-variance control variates can significantly reduce the variance of pathwise gradient estimators without requiring complex assumptions.
We propose a training method for deep neural network (DNN)-based source enhancement to increase objective sound quality assessment (OSQA) scores such as the perceptual evaluation of speech quality (PESQ). In many conventional studies, DNNs have been used as a mapping function to estimate time-frequency masks and traine…
VarGrad reduces variance in ELBO gradient estimation for variational inference.
problem Improving the variance of gradient estimators in variational inference.
method VarGrad uses a new log-variance loss to estimate the ELBO gradient, achieving lower variance than the score function method.
result VarGrad offers a lower variance gradient estimator compared to other methods.
Directly estimates Fisher score for likelihood maximization.
problem Intractable likelihood functions with model simulations.
method Gradient-based optimization using local score matching and linear parameterization.
result Efficient approximation of Fisher score improves likelihood maximization.
Quantized Variational Inference improves ELBO optimization with fast convergence.
problem Maximizing Evidence Lower Bound (ELBO) for variational inference.
method Optimal Voronoi Tesselation for variance-free gradients, Richardson extrapolation for asymptotic improvement.
result Quantized Variational Inference leads to fast convergence with comparable computational cost.
The reparameterization trick is widely used in variational inference as it yields more accurate estimates of the gradient of the variational objective than alternative approaches such as the score function method. Although there is overwhelming empirical evidence in the literature showing its success, there is relative…
Density estimation is a fundamental problem in statistical learning. This problem is especially challenging for complex high-dimensional data due to the curse of dimensionality. A promising solution to this problem is given here in an inference-free hierarchical framework that is built on score matching. We revisit the…
Unified view of score estimators for flexible densities.
problem Estimating the score from unknown distributions.
method Regularized nonparametric regression framework.
result Unified convergence analysis and new estimators with desirable properties.
New cross-validation methods for Gaussian process regression with efficient gradient computation.
problem Estimating parameters of Gaussian process covariance functions.
method Derive new cross-validation criteria and efficient adjoint computation of gradients.
result Efficient method for evaluating cross-validation criteria and their gradients.
Semiparametric method removes bias in functional bilevel gradient estimation.
problem First-order bias in plug-in hypergradient when lower-level problem is nonparametric.
method Semiparametric debiasing theory based on efficient influence function leads to cross-fitted orthogonal hypergradient estimator.
result Asymptotic normality and uniform control over outer parameter established for the estimator.
Kernel SVGD improves high-dimensional inference with noise adaptation.
problem Challenges in high-dimensional inference with SVGD.
method Noise Conditional Kernel SVGD (NCK-SVGD) with entropic regularization.
result NCK-SVGD produces samples comparable to GANs and SGLD on computer vision benchmarks.
Two new estimators improve VAE training for hierarchical and prior parameters.
problem Efficient gradient estimation for VAEs with hierarchical and prior parameters.
method Developed two generalizations of Doubly-Reparameterized Gradient Estimators (DReGs) for VAEs.
result Improved training of conditional and hierarchical VAEs on image modeling tasks.
We propose a new model for supervised learning to rank. In our model, the relevance labels are assumed to follow a categorical distribution whose probabilities are constructed based on a scoring function. We optimize the training objective with respect to the multivariate categorical variables with an unbiased and low-…
We propose a simple and general variant of the standard reparameterized gradient estimator for the variational evidence lower bound. Specifically, we remove a part of the total derivative with respect to the variational parameters that corresponds to the score function. Removing this term produces an unbiased gradient …
Heuristic weighting improves denoising score matching without requiring noise distribution assumptions.
problem Improving denoising score matching without assuming noise distribution.
method Demonstrated heteroskedasticity, derived optimal weighting functions, and provided theoretical and empirical comparisons.
result Heuristical weighting function can achieve lower variance than optimal weighting, facilitating more stable and efficient training.
Generative method avoids function estimation for data generation.
problem Challenges in function estimation for generative models.
method Deterministic point transport with gradient descent.
result Data generation possible without function estimation.
Study proposes a differentiable surrogate loss function for optimizing Fβ score in binary classification with imbalanced data.
problem Non-differentiability of Fβ score makes it unsuitable for optimization by gradient-based learning. method Investigated relationship between Fβ score and loss functions, proposed a differentiable surrogate loss function. result Gradient paths of the proposed surrogate Fβ loss function approximate the gradient paths of the Fβ score. New geometric analysis shows L2 score error is flawed for diffusion models.
problem Score matching errors in diffusion models do not fully capture distributional quality.
method Decomposed score errors into gradient and solenoidal components, focusing on gradient's role in Fokker-Planck dynamics.
result Only gradient component affects marginal distributional quality; solenoidal component is structurally invisible.
Score matching errors are not sufficient for measuring diffusion model quality.
problem The L2 score matching error is not a reliable measure of diffusion model performance. method Decomposed score errors into gradient and solenoidal components and analyzed their geometric properties.
result Only the gradient component of the score error affects the marginal distributional quality.
Maximum likelihood estimation of energy-based models is a challenging problem due to the intractability of the log-likelihood gradient. In this work, we propose learning both the energy function and an amortized approximate sampling mechanism using a neural generator network, which provides an efficient approximation o…
New framework for neural network score estimation in diffusion models.
problem Rigorous guarantees for practical score estimation with neural networks.
method Developed a mathematical framework for score estimation with GD-trained neural networks, addressing optimization and generalization.
result Established minimax-optimal generalization bounds for GD-trained neural networks in diffusion models.
BoXHED boosts hazard estimation for dynamic health risk scores.
problem Analyzing time-varying health vitals for disease onset prediction.
method Gradient boosting for nonparametric hazard function estimation with time-dependent covariates.
result Novel interaction effects among risk factors identified in cardiovascular disease onset data.
Paper develops unbiased gradient estimator for continuous-time models.
problem Estimating unbiased gradient of log-likelihood for continuous-time models.
method Doubly randomized scheme with coupled conditional particle filter (CCPF).
result Unbiased gradient estimate facilitates gradient-based algorithms.
Noise-corrected Langevin algorithm improves sampling from noisy data.
problem Sampling from noisy data with biased score function.
method Noise-corrected Langevin algorithm using noisy score function.
result Bias due to noisy data is removed, improving sampling accuracy.
Paper introduces a new gradient estimator for SNNs.
problem High variance in score function gradient estimator impedes SNNs training.
method Developed a differentiable point process to derive path-wise gradient estimator.
result Demonstrated effectiveness of path-wise gradient estimator through simulations.
New algorithms improve uncertainty estimation in satellite precipitation predictions.
problem Lack of uncertainty estimates in machine learning spatial precipitation predictions from satellite data.
method Benchmarked six algorithms including LightGBM, compared using quantile scoring functions and rules.
result LightGBM outperformed other algorithms in quantile scoring rule by 11.10%.
Significant advances have been made recently on training neural networks, where the main challenge is in solving an optimization problem with abundant critical points. However, existing approaches to address this issue crucially rely on a restrictive assumption: the training data is drawn from a Gaussian distribution. …
The paper interprets diffusion models as gradient descent and proposes a new sampler.
problem Improving the efficiency and quality of diffusion models.
method Interprets diffusion models as gradient descent and proposes a new sampler.
result The new sampler achieves state-of-the-art FID scores and generates high quality samples.
Proposes sigmoidF1 loss for multilabel classification, improving performance metrics.
problem Lack of smooth, tractable loss functions for multilabel classification.
method Introduces sigmoidF1, a smooth F1 score surrogate loss function.
result sigmoidF1 outperforms other loss functions on various datasets and metrics.
Optimal score function estimation via empirical risk minimization
problem Estimating the score function of a probability measure on the flat torus from a sample
method Constraining the hypothesis space to a Sobolev ball
result Minimax estimation rates are achieved
Learning models with discrete latent variables using stochastic gradient descent remains a challenge due to the high variance of gradient estimates. Modern variance reduction techniques mostly consider categorical distributions and have limited applicability when the number of possible outcomes becomes large. In this w…
Storchastic improves stochastic AD for complex models in RL and VI.
problem Handling intractable expectations in RL and VI.
method Introduces Storchastic, a framework for AD of stochastic computation graphs with various gradient estimation methods.
result Provable unbiasedness and variance reduction for higher-order gradients.
A method makes particle filters differentiable without altering their forward pass.
problem Compatibility issues between particle filters and automatic differentiation.
method Introduces a correction to particle weights using the stop-gradient operator.
result Automatic differentiation produces good estimators for gradients and second-order derivatives.
Proposes MIGE for accurate MI gradient estimation in high-dimensional settings.
problem Intractability of MI in continuous and high-dimensional settings.
method Score estimation of implicit distributions for gradient estimation of MI.
result MIGE provides tight and smooth gradient estimation of MI in high-dimensional settings.
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…
This paper is a broad and accessible survey of the methods we have at our disposal for Monte Carlo gradient estimation in machine learning and across the statistical sciences: the problem of computing the gradient of an expectation of a function with respect to parameters defining the distribution that is integrated; t…
A diffusion model estimates data manifold dimension by tracking likelihood increases.
problem Estimating the intrinsic dimension of data manifolds.
method Trained diffusion model approximates score function, revealing manifold directionality.
result Diffusion model provides an approximation of the tangent space's dimension.
Langevin dynamics fails to produce accurate samples even with small score function errors.
problem Robustness of Langevin dynamics to score function errors.
method Analysis of Langevin dynamics and score function errors.
result Langevin dynamics produces a distribution far from the target distribution in TV distance even with small L2 errors in the score function. New polynomial convergence guarantees for SGM on general data distributions.
problem Efficient guarantees for multimodal and non-smooth distributions in SGM.
method Polynomial convergence guarantees for denoising diffusion models on general data distributions, with no assumptions on functional inequalities or smoothness.
result Wasserstein distance guarantees for distributions of bounded support or decaying tails, and TV guarantees for further smoothness assumptions.
ScoreStop uses gradient tests to stop gradient boosting early.
problem Overfitting in gradient boosted decision trees.
method ScoreStop uses a functional score test based on gradients to stop boosting.
result ScoreStop is competitive with loss-based early stopping methods.