archive
Every paper Pith has read. Search by title, abstract, or pith.
883 papers in math.ST · page 11
-
Common minimal-sufficiency test fails in general
Version-Robust Methods for Identifying Minimal Sufficient Statistics
-
Analyst-oracle gap sets limit on recovering latent genealogies from sequences
Sequential learning theory for Markov genealogy processes
-
Wasserstein projection makes shape-constrained density estimation convex
Shape-constrained density estimation with Wasserstein projection
-
Laguerre series yield minimax optimal rates for varying coefficients
Minimax estimation for Varying Coefficient Model via Laguerre Series
-
Weighted Chernoff information governs optimal loss exponent
Weighted Chernoff information and optimal loss exponent in context-sensitive hypothesis testing
-
Artificial Replay halves interactions to compare bandit policies
Design Experiments to Compare Multi-armed Bandit Algorithms
-
Diffusion models converge at rates set by data's intrinsic dimension
Generalization Properties of Score-matching Diffusion Models for Intrinsically Low-dimensional Data
-
Strong graph disjointness holds exactly when one graph is a tree
Graph Disjointness with Applications to Reversible Markov Chains
-
Safe reference policy sets limits on how far a new policy may stray
Conformal Policy Control
-
Flow matching converges at intrinsic-dimension rates on manifolds
Flow Matching is Adaptive to Manifold Structures
-
Spike extrapolation classifies synapses from neuron pairs alone
Synaptic Classification via Spike-Triggered Extrapolation
-
Weakly convex-concave penalty recovers sparse signals from quadratic data
Support Recovery and $\ell_2$-Error Bound for Sparse Regression with Quadratic Measurements via Weakly-Convex-Concave Regularization
-
Sequential model access boosts membership inference power
Sequential Membership Inference Attacks
-
Gamma populations remove all finite-sample bias from order-statistic inequality estimators
Bias analysis of a linear order-statistic inequality index estimator: Unbiasedness under gamma populations
-
Subexponential noise yields poly-log log-concave sampling accuracy
High-accuracy log-concave sampling with stochastic queries
-
Dickman distribution extends to vectors with preserved divisibility
Multidimensional Dickman distribution and operator selfdecomposability
-
Poly-time algorithm recovers regressor from unknown-truncated data
Linear Regression with Unknown Truncation Beyond Gaussian Features
-
20/80 rule emerges structurally from truncated normal and exponential data
Formalization of the generalized Pareto principle and structural typicality of the 20/80-rule
-
Tight detection thresholds derived for masked bipartite latent graphs
Information-Theoretic Thresholds for Bipartite Latent-Space Graphs under Noisy Observations
-
Föllmer process selected to minimize path KL error
Variational Optimality of F\"ollmer Processes in Generative Diffusions
-
Betting strategies go bankrupt almost surely under the null
Almost sure null bankruptcy of testing-by-betting strategies
-
Good-Toulmin estimator is unique for small-m unseen species prediction
The Unseen Species Problem Revisited
-
Bayes theorem equals the unique optimal data-deletion rule
Optimal information deletion and Bayes' theorem
-
Regression estimators converge weakly in Hilbert space
High-dimensional linear regression inference via $\ell^2$ weak convergence
-
Ellipticity makes continuous-time RL as easy as supervised learning
Continuous-time reinforcement learning: ellipticity enables model-free value function approximation
-
Blurred distance permits assumption-free two-sample testing
Distribution-free two-sample testing with blurred total variation distance
-
Weighted risk minimization yields fast rates under distribution drift
Fast Rates for Nonstationary Weighted Risk Minimization
-
Regularized SVGD delivers finite-particle convergence rates
Finite-Particle Rates for Regularized Stein Variational Gradient Descent
-
Neural drift estimates classify diffusion trajectories
Plug-In Classification of Drift Functions in Diffusion Processes Using Neural Networks
-
Diffusion samplers reach δ accuracy in polylog steps
High-accuracy sampling for diffusion models and log-concave distributions
-
SHAP attributions spot complementary anomaly detectors
Analyzing Shapley Additive Explanations to Understand Anomaly Detection Algorithm Behaviors and Their Complementarity
-
Quantum tomography gains valid confidence sets at any stopping time
Anytime-Valid Quantum State Tomography via Confidence Sequences
-
Latent-IMH draws exact posterior samples using cheap operator approximations
Latent-IMH: Efficient Bayesian Inference for Inverse Problems with Approximate Operators
-
P-P process converges in L1 iff its curve is absolutely continuous
Convergence in distribution of the P-P process in $L^1[0,1]$
-
Szász-Mirakyan CDF estimator cuts variance inside orthant but not at boundary
Asymptotic properties of the multivariate Sz\'{a}sz-Mirakyan estimator for cumulative distribution functions on the nonnegative orthant
-
Extreme scores in large equal tournaments converge to known limits
Extreme Score Distributions in Countable-Outcome Round-Robin Tournaments of Equally Strong Players
-
Transformers achieve optimal nonparametric regression rates with log n parameters
Efficient and Minimax Optimal In-context Nonparametric Regression with Transformers
-
Unilateral shift unifies invertibility with algebraic transfer functions
Stationary Process Invertibility and the Unilateral Shift Operator
-
Quantile-winsorized max-test matches standard power
Robustness for free: asymptotic size and power of max-tests in high dimensions
-
Hypotheses admit consistent finite-precision tests iff both are F_sigma
Topological Criteria for Hypothesis Testing with Finite-Precision Measurements
-
PCA sign rule bounds error in non-isotropic mixture clustering
On spectral clustering under non-isotropic Gaussian mixture models
-
Refinement beats Lasso prediction error in some tuning regimes
Prediction Suboptimality of the Lasso in Sparse Linear Regression
-
Matching bounds fix sample complexity for composite quantum tests
Sample Complexity of Composite Quantum Hypothesis Testing
-
Estimator keeps invariance guarantees despite missing outcomes
Multi-environment Invariance Learning with Missing Data
-
Phases in ERGMs satisfy approximate FKG inequality
Approximate FKG inequalities for phase-bound spin systems, with applications to central limit theorems for exponential random graphs
-
Lossless compression sample complexity set by Rényi entropy of order 1/2
The Sample Complexity of Lossless Data Compression
-
Skeleton matrix fixes recurrent classes and periods of higher-order chains
Reduction and classification of higher-order Markov chains
-
Multicalibration needs Ω(T^{2/3}) error online
Optimal Lower Bounds for Online Multicalibration
-
Instrumental variables recover local dose-response curves
Double Machine Learning of Continuous Treatment Effects with General Instrumental Variables
-
Spectral functional detects causality at linear sample scaling
Order-Constrained Spectral Causality for Multivariate Time Series