analyze score matching

This work addresses the challenge of parameter inference and uncertainty quantification in generative models where the likelihood function is intractable. We propose a likelihood-free inference framework that approximates the gradient of the log-likelihood via structured score matching, enabling efficient estimation through gradient-based optimization combined with bootstrap resampling. The key innovation lies in a regularized neural network architecture that embeds statistical structure and a tailored score matching estimator, which together ensure theoretical convergence while substantially improving estimation accuracy and scalability. Numerical experiments demonstrate that the proposed method outperforms existing approaches in both parameter estimation accuracy and the reliability of uncertainty quantification.

Maximum likelihood estimation (MLE) for finite point processes—random point configurations on bounded domains—is computationally prohibitive, while existing score-matching methods lack rigorous theoretical foundations. Method: We establish the first score-matching framework grounded in the Janossy measure and propose an autoregressive weighted score-matching estimator. Theoretically, we expose an intrinsic identifiability limitation of nonparametric score matching—its inability to uniquely recover the true distribution—and introduce a survival-classification augmentation strategy to construct a differentiable training objective that avoids both integration and normalization constants. Our approach unifies Janossy measures, weighted score matching, autoregressive modeling, and survival loss, ensuring statistical consistency while drastically improving computational efficiency. Results: Experiments on synthetic and real spatiotemporal point process data demonstrate that our method accurately recovers intensity functions with accuracy comparable to MLE, while achieving significantly faster training.

This work addresses the challenge of score estimation in the presence of latent variables, where conventional denoising score matching (DSM) suffers from high variance at low noise levels, and target score matching (TSM) is inapplicable due to the unavailability of clean-data scores. To overcome this limitation, the authors propose Latent-variable Target Score Matching (LTSM), which extends TSM to settings with latent variables for the first time. LTSM leverages the score of the joint distribution to provide low-variance supervision for the marginal score and integrates DSM into a hybrid training strategy that ensures robustness across varying noise scales. Experimental results demonstrate that LTSM substantially reduces estimation variance, leading to improved score estimation accuracy and enhanced generative sample quality.

This paper addresses the design of proper scoring rules for multidimensional forecasting settings, aiming to incentivize forecasters to exert effort and truthfully report their beliefs. Methodologically, it introduces the first optimization framework explicitly targeting *effort incentives*, integrating game-theoretic modeling with convex optimization. For simple settings, it derives closed-form characterizations of optimal rules; for general cases, it develops an efficient and exact algorithm; and it identifies several structurally simple approximate rules with near-optimal performance. Theoretical analysis reveals that classical proper scoring rules—such as the quadratic score—can substantially deviate from optimality under multidimensional effort. In contrast, the proposed algorithm computes exact optimal rules, while the simple approximations achieve over 95% of the optimal incentive efficiency. These results establish a new paradigm for information design and prediction market mechanisms, bridging incentive alignment with practical implementability.

This work addresses the challenge of selecting weighting functions in generalized score matching, which critically affects estimation efficiency. By embedding the method within the generalized method of moments (GMM) framework, the paper establishes its equivalence to Stein moment estimation. Through an extension of the Stein class to the generalized moment setting, a novel class of estimators is constructed, offering theoretically guaranteed superior statistical properties. This approach not only circumvents the subjectivity inherent in traditional tuning of weighting functions but also achieves optimal estimation efficiency, thereby significantly enhancing both the practical applicability and theoretical rigor of generalized score matching.

This study addresses the evaluation and selection of source-level likelihood ratio (LR) systems for forensic evidence-to-reference comparison tasks by proposing an integrated analytical framework that balances performance and practical feasibility. The authors employ strictly proper scoring rules to quantify how effectively each system updates Bayesian prior odds and present the first systematic comparison among specific-source feature-based, common-source anchored, and unanchored score-based LR approaches. Their findings reveal that specific-source feature-based LRs achieve the highest performance but incur substantial experimental costs, whereas common-source feature-based methods offer strong discriminative power with significantly reduced implementation complexity. All LR systems substantially outperform a baseline relying solely on prior odds. This work thus provides both theoretical grounding and practical guidance for selecting LR systems in forensic practice.

This work addresses the lack of non-asymptotic sample complexity guarantees for learning high-dimensional continuous exponential family distributions, particularly in the challenging setting of unbounded support. By employing score matching to perform structure learning on polynomial-form exponential family models, the study establishes the first finite-sample error bounds for this class of models through a synthesis of high-dimensional probabilistic modeling and non-asymptotic statistical analysis. The results demonstrate that the required sample size scales polynomially with the ambient dimension, thereby providing the first rigorous sample complexity guarantee for learning high-dimensional exponential family distributions with unbounded support. This contribution fills a critical theoretical gap in the non-asymptotic understanding of continuous exponential families.

This paper investigates gradient descent dynamics for overparameterized score matching in learning Gaussian and Gaussian mixture distributions. The theoretical analysis centers on how noise scale, model overparameterization, and initialization strategies affect optimization behavior. Our contributions are threefold: (1) We establish, for the first time within the score matching framework, global convergence of gradient descent for Gaussian mixture models with at least three components; (2) We characterize the critical role of noise scale—proving global convergence under large noise, while deriving precise convergence conditions and constructing explicit divergence counterexamples in the low-noise regime; (3) We show that small initialization guarantees parameter convergence, whereas random initialization—though causing divergence of some parameters—yields loss convergence at rate $1/ au$ almost surely, with a matching lower bound. Collectively, these results provide a unified characterization of multiscale optimization dynamics and implicit generalization bias.

This work establishes a rigorous theoretical connection between drift models and score-based generative models through the lens of score matching under kernel-smoothed distributions. It demonstrates, for the first time, that the drift field induced by a Gaussian kernel is equivalent to the difference between the scores of the smoothed data distribution and the model distribution—a result extended to general radial kernels. By integrating Tweedie’s formula, kernel density estimation, and high-dimensional probabilistic analysis, the study further reveals an intrinsic link to the Distributional Matching Distillation (DMD) framework. Moreover, it provides sharp error bounds for the Laplace kernel in low-temperature and high-dimensional regimes, proving its ability to accurately approximate the score-matching objective under these challenging conditions.