MI
Masaaki Imaizumi
cs.LGstat.MLmath.STstat.THstat.MEecon.EMcs.AIcond-mat.dis-nnmath.OCMachine Learning
On Valency
published · living versionsW_3p6ka57n·v1 · currentpublished
High-Dimensional Off-Policy Evaluation without Sample Splitting using an Onsager Correction
with Saksham Jain, Guillaume Braun
1 version
Preprints & journals
71 papers in the corpus · 2015–2026Fast Learning Rate Transfer in Shallow Linear Networks at Growing Training Horizons2609.35029v1 · Mana Sakai, Masaaki Imaizumi2026 · 0 citationsarXiv
Overfitting of Spectral Gradient Descent: How Matrix Geometry shapes Generalization and Implicit Bias2609.32270v1 · Guillaume Braun, Ichiro Hashimoto, Masaaki Imaizumi2026 · 0 citationsarXiv
Anti Mode-Collapse in Mean-Field Transformer via Auxiliary Variables2605.30229v1 · Masaaki Imaizumi, Masanori Koyama, Noboru Isobe et al.2026 · 0 citationsarXiv
Minimax Optimal Estimation of Transport-Growth Pairs in Unbalanced Optimal Transport2605.08705v1 · Donlapark Ponnoprat, Noboru Isobe, Masaaki Imaizumi2026 · 0 citationsarXiv
Training-Induced Escape from Token Clustering in a Mean-Field Formulation of Transformers2605.07772v2 · Noboru Isobe, Daisuke Inoue, Masaaki Imaizumi2026 · 0 citationsarXiv
CITE: Anytime-Valid Statistical Inference in LLM Self-Consistency2605.05873v1 · Hirofumi Ota, Naoto Iwase, Yuki Ichihara et al.2026 · 0 citationsarXiv
Why Mean Pooling Works: Quantifying Second-Order Collapse in Text Embeddings2604.27398v1 · Tomomasa Hara, Hiroto Kurita, Masaaki Imaizumi et al.2026 · 0 citationsarXiv
Dichotomy of Feature Learning and Unlearning: Fast-Slow Analysis on Neural Networks with Stochastic Gradient Descent2602.07378v1 · Shota Imai, Sota Nishiyama, Masaaki Imaizumi2026 · 0 citationsarXiv
High-Dimensional Limit of Stochastic Gradient Flow via Dynamical Mean-Field Theory2602.06320v2 · Sota Nishiyama, Masaaki Imaizumi2026 · 0 citationsarXiv
Spectral Gradient Descent Mitigates Anisotropy-Driven Misalignment: A Case Study in Phase Retrieval2601.22652v2 · Guillaume Braun, Han Bao, Wei Huang et al.2026 · 0 citationsarXiv
Neuron Block Dynamics for XOR Classification with Zero-Margin2602.00172v1 · Guillaume Braun, Masaaki Imaizumi2026 · 0 citationsProceedings of the 29th International Conference on Artificial Intelligence and Statistics (AISTATS) 2026
Finite-Sample Inference for Sparsely Permuted Linear Regression2601.14872v2 · Hirofumi Ota, Masaaki Imaizumi2026 · 0 citationsarXiv
Demystifying MaskGIT Sampler and Beyond: Adaptive Order Selection in Masked Diffusion2510.04525v3 · Satoshi Hayakawa, Yuhta Takida, Masaaki Imaizumi et al.2025 · 0 citationsTransactions on Machine Learning Research, 2026
SONA: Learning Conditional, Unconditional, and Mismatching-Aware Discriminator2510.04576v1 · Yuhta Takida, Satoshi Hayakawa, Takashi Shibuya et al.2025 · 0 citationsarXiv
Precise Dynamics of Diagonal Linear Networks: A Unifying Analysis by Dynamical Mean-Field Theory2510.01930v2 · Sota Nishiyama, Masaaki Imaizumi2025 · 0 citationsarXiv
Optimal Dynamic Regret by Transformers for Non-Stationary Reinforcement Learning2508.16027v2 · Baiyuan Chen, Shinji Ito, Masaaki Imaizumi2025 · 0 citationsNeurIPS 2025
Implications From the Analogous Relationship Between Evolutionary and Learning Processes.40484809 · Leong, Jason Cheok Kuan, Imaizumi, Masaaki, Innan, Hideki et al.2025 · 0 citationsBioEssays : news and reviews in molecular, cellular and developmental biology. 2025;47(8):e70027
Infinite-Width Limit of a Single Attention Layer: Analysis via Tensor Programs2506.00846v3 · Mana Sakai, Ryo Karakida, Masaaki Imaizumi2025 · 0 citationsAdvances in Neural Information Processing Systems 38 (2025), 40079-40113
High-dimensional Nonparametric Contextual Bandit Problem2505.14102v1 · Shogo Iwazaki, Junpei Komiyama, Masaaki Imaizumi2025 · 0 citationsarXiv
Minimax Rates of Estimation for Optimal Transport Map between Infinite-Dimensional Spaces2505.13570v3 · Donlapark Ponnoprat, Masaaki Imaizumi2025 · 0 citationsarXiv
Precise gradient descent training dynamics for finite-width multi-layer neural networks2505.04898v1 · Qiyang Han, Masaaki Imaizumi2025 · 0 citationsarXiv
Learning a Single Index Model from Anisotropic Data with vanilla Stochastic Gradient Descent2503.23642v1 · Guillaume Braun, Minh Ha Quang, Masaaki Imaizumi2025 · 0 citationsProceedings of the 28th International Conference on Artificial Intelligence and Statistics (AISTATS) 2025
Landscape computations for the edge of chaos in nonlinear dynamical systems2503.06393v1 · Motoki Nakata, Masaaki Imaizumi2025 · 0 citationsJSIAM Letters vol. 17, 65 (2025)
Approximation of Permutation Invariant Polynomials by Transformers: Efficient Construction in Column-Size2502.11467v1 · Naoki Takeshita, Masaaki Imaizumi2025 · 0 citationsarXiv
Federated Learning with Relative Fairness2411.01161v1 · Shogo Nakakita, Tatsuya Kaneko, Shinya Takamaeda-Yamazaki et al.2024 · 0 citationsarXiv
Universality of estimators for high-dimensional linear models with block dependency2410.19244v2 · Toshiki Tsuda, Masaaki Imaizumi2024 · 1 citationarXiv
Distillation of Discrete Diffusion through Dimensional Correlations2410.08709v4 · Satoshi Hayakawa, Yuhta Takida, Masaaki Imaizumi et al.2024 · 0 citationsarXiv
Wasserstein -Centers Clustering for Distributional Data2407.08228v4 · Ryo Okano, Masaaki Imaizumi2024 · 3 citationsarXiv
Effect of Random Learning Rate: Theoretical Analysis of SGD Dynamics in Non-Convex Optimization via Stationary Distribution2406.16032v2 · Naoki Yoshida, Shogo Nakakita, Masaaki Imaizumi2024 · 0 citationsTransactions on Machine Learning Research, 2835-8856, 2025
Automatic Domain Adaptation by Transformers in In-Context Learning2405.16819v1 · Ryuichiro Hataya, Kota Matsui, Masaaki Imaizumi2024 · 0 citationsarXiv
High-Dimensional Single-Index Models: Link Estimation and Marginal Inference2404.17812v2 · Kazuma Sawaya, Yoshimasa Uematsu, Masaaki Imaizumi2024 · 0 citationsarXiv
Effect of Weight Quantization on Learning Models by Typical Case Analysis2401.17269v1 · Shuhei Kashiwamura, Ayaka Sakata, Masaaki Imaizumi2024 · 2 citationsarXiv
CATE Lasso: Conditional Average Treatment Effect Estimation with High-Dimensional Linear Regression2310.16819v1 · Masahiro Kato, Masaaki Imaizumi2023 · 3 citationsarXiv
Uniform Confidence Band for Optimal Transport Map on One-Dimensional Data2307.09257v2 · Donlapark Ponnoprat, Ryo Okano, Masaaki Imaizumi2023 · 1 citationElectronic Journal of Statistics 2024, Vol. 18, No. 1, 515-552
Distribution-on-Distribution Regression with Wasserstein Metric: Multivariate Gaussian Case2307.06137v3 · Ryo Okano, Masaaki Imaizumi2023 · 8 citationsarXiv
Sup-Norm Convergence of Deep Neural Network Estimator for Nonparametric Regression by Adversarial Training2307.04042v1 · Masaaki Imaizumi2023 · 3 citationsarXiv
High-dimensional Contextual Bandit Problem without Sparsity2306.11017v2 · Junpei Komiyama, Masaaki Imaizumi2023 · 0 citationsAdvances in Neural Information Processing Systems, 36, 49416-49427, 2023
Moment-Based Adjustments of Statistical Inference in High-Dimensional Generalized Linear Models2305.17731v5 · Kazuma Sawaya, Yoshimasa Uematsu, Masaaki Imaizumi2023 · 0 citationsarXiv
Bayesian Analysis for Over-parameterized Linear Model via Effective Spectra2305.15754v3 · Tomoya Wakayama, Masaaki Imaizumi2023 · 0 citationsarXiv
Benign Overfitting of Non-Sparse High-Dimensional Linear Regression with Correlated Noise2304.04037v3 · Toshiki Tsuda, Masaaki Imaizumi2023 · 3 citationsarXiv
Asymptotically Optimal Fixed-Budget Best Arm Identification with Variance-Dependent Bounds2302.02988v2 · Masahiro Kato, Masaaki Imaizumi, Takuya Ishihara et al.2023 · 1 citationarXiv
SAN: Inducing Metrizability of GAN with Discriminative Normalized Linear Layer2301.12811v4 · Yuhta Takida, Masaaki Imaizumi, Takashi Shibuya et al.2023 · 8 citationsarXiv
Dimension-free Bounds for Sum of Dependent Matrices and Operators with Heavy-Tailed Distribution2210.09756v3 · Shogo Nakakita, Pierre Alquier, Masaaki Imaizumi2022 · 0 citationsElectron. J. Statist. 18(1): 1130-1159 (2024)
Best Arm Identification with Contextual Information under a Small Gap2209.07330v4 · Masahiro Kato, Masaaki Imaizumi, Takuya Ishihara et al.2022 · 0 citationsarXiv
Hypothesis Test and Confidence Analysis With Wasserstein Distance on General Dimension.35534006 · Imaizumi, Masaaki, Ota, Hirofumi, Hamaguchi, Takuo2022 · 10 citationsNeural computation. 2022;34(6):1448-1487
Benign Overfitting in Time Series Linear Models with Over-Parameterization2204.08369v3 · Shogo Nakakita, Masaaki Imaizumi2022 · 1 citationarXiv
Inference for Projection-Based Wasserstein Distances on Finite Spaces2202.05495v1 · Ryo Okano, Masaaki Imaizumi2022 · 5 citationsarXiv
Benign-Overfitting in Conditional Average Treatment Effect Prediction with Linear Regression2202.05245v2 · Masahiro Kato, Masaaki Imaizumi2022 · 3 citationsarXiv
Unified Perspective on Probability Divergence via Maximum Likelihood Density Ratio Estimation: Bridging KL-Divergence and Integral Probability Metrics2201.13127v1 · Masahiro Kato, Masaaki Imaizumi, Kentaro Minami2022 · 1 citationarXiv
Optimal Best Arm Identification in Two-Armed Bandits with a Fixed Budget under a Small Gap2201.04469v8 · Masahiro Kato, Kaito Ariu, Masaaki Imaizumi et al.2022 · 0 citationsarXiv
Career total: 94 works. 71 are in this corpus.Showing the 50 most recent.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.