Statistics (Machine Learning) Papers

Semiparametric Efficient Bilevel Gradient Estimation Fares El Khoury, Houssam Zenati, Nathan Kallus, Michael Arbel, Aurélien Bibaut arxiv.org/abs/2605.21341 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Large-Step Training Dynamics of a Two-Factor Linear Transformer Model Krishnakumar Balasubramanian arxiv.org/abs/2605.21292 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙰𝙸 𝚌𝚜.𝙻𝙶 𝚖𝚊𝚝𝚑.𝙳𝚂]

Theoretical guidelines for annealed Langevin dynamics in compositional simulation-based inference Camille Touron, Gabriel V. Cardoso, Julyan Arbel, Pedro L. C. Rodrigues arxiv.org/abs/2605.21253 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Federated LoRA Fine-Tuning for LLMs via Collaborative Alignment Shuaida He, Liwen Chen, Long Feng arxiv.org/abs/2605.21217 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

A Rigorous, Tractable Measure of Model Complexity Oskar Allerbo, Thomas B. Schön arxiv.org/abs/2605.21167 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Conditioning Gaussian Processes on Almost Anything Henry Moss, Lachlan Astfalck, Thomas Cowperthwaite, Colin Doumont, Sam Willis, Philipp Hennig, Christopher Nemeth, Andrew Zammit-Mangion arxiv.org/abs/2605.21041 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶 𝚜𝚝𝚊𝚝.𝙼𝙴]

Group-Aware Matrix Estimation and Latent Subspace Recovery Hamza Golubovic, Matthew Shen, Genevera I. Allen, Tarek M. Zikry arxiv.org/abs/2605.20559 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶 𝚜𝚝𝚊𝚝.𝙰𝙿 𝚜𝚝𝚊𝚝.𝙼𝙴]

Spectral bandits for smooth graph functions with applications in recommender systems Tomáš Kocák, Michal Valko, Rémi Munos, Branislav Kveton, Shipra Agrawal arxiv.org/abs/2605.20552 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Sample Complexity of Transfer Learning: An Optimal Transport Approach Haoyang Cao, Xin Guo, Wenpin Tang, Guan Wang arxiv.org/abs/2605.20545 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Contradiction Graphs Determine VC Dimension Jesse Campbell, Daniel Ibaibarriaga, Lev Reyzin arxiv.org/abs/2605.20434 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙳𝙼 𝚌𝚜.𝙻𝙶]

Corrected Integrated Laplace Approximation for Bayesian Inference in Latent Gaussian Models Jinlin Lai, Charles C. Margossian, Daniel R. Sheldon arxiv.org/abs/2605.20345 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Multi-Head Attention as Ensemble Nadaraya-Watson Estimation: Variance Reduction, Decorrelation, and Optimal Head Diversity Ernest Fokoué arxiv.org/abs/2605.20271 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Goal-Oriented Lower-Tail Calibration of Gaussian Processes for Bayesian Optimization Aurélien Pion, Emmanuel Vazquez arxiv.org/abs/2605.20145 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶 𝚜𝚝𝚊𝚝.𝙼𝙴]

Optimizing Computational-Statistical Runtime for Wasserstein Distance Estimation Peter Matthew Jacobs, Jeff M. Phillips arxiv.org/abs/2605.20122 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙲𝙲 𝚌𝚜.𝙻𝙶]

Tail Annealing for Heavy-Tailed Flow Matching Jean Pachebat arxiv.org/abs/2605.20068 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Probabilistic Multivariate Time Series Forecasting with Diffusion Copulas David Huk, Dongshan Wang, Miha Bresar arxiv.org/abs/2605.19685 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Increasing Missingness to Reduce Bias: Richardson-SGD with Missing Data Ferdinand Genans, Erwan Scornet arxiv.org/abs/2605.19641 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Gaussian Approximation and Multiplier Bootstrap for Federated Linear Stochastic Approximation Ilya Levin, Maksim Shuklin, Eric Moulines, Paul Mangold, Sergey Samsonov arxiv.org/abs/2605.19629 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶 𝚖𝚊𝚝𝚑.𝙾𝙲]

Posterior Contraction of Lévy Adaptive B-spline Regression in Besov Spaces Jeunghun Oh, Sewon Park, Jaeyong Lee arxiv.org/abs/2605.19610 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]

Density-Ratio Losses for Post-Hoc Learning to Defer Alexander Soen, Ragnar Thobaben, Joakim Jaldén, Richard Nock arxiv.org/abs/2605.19557 [𝚜𝚝𝚊𝚝.𝙼𝙻 𝚌𝚜.𝙻𝙶]