Papers
Published Papers
A Decision-Theoretic View of Test-Time Training: When, How Far, and Which Directions to Adapt
Tomoya Wakayama
Proceedings of the 43rd International Conference on Machine Learning, 306, 124279–124317 (2026)
A decision-theoretic analysis of test-time training, including adaptation distance and direction selection.
In-Context Learning Is Provably Bayesian Inference: A Generalization Theory for Meta-Learning
Tomoya Wakayama, Taiji Suzuki
Proceedings of the 43rd International Conference on Machine Learning, 306, 124318–124351 (2026)
A generalization theory showing when in-context learning implements Bayesian inference in meta-learning.
Similarity-based Random Partition Distribution for Clustering Functional Data
Tomoya Wakayama, Shonosuke Sugasawa, Genya Kobayashi
Journal of the Royal Statistical Society Series C: Applied Statistics, 75(1), 100–119 (2026)
A nonparametric Bayesian clustering method for functional data that incorporates similarity information.
Reconciling Functional Data Regression with Excess Bases
Tomoya Wakayama, Hidetoshi Matsui
Behaviormetrika, 53(2), 551–569 (2026)
A study of overparameterization and double-descent phenomena in functional data regression with basis expansions.
Ensemble Prediction via Covariate-dependent Stacking
Tomoya Wakayama, Shonosuke Sugasawa
Statistics and Computing, 35, 212 (2025)
A covariate-dependent stacking framework for ensemble prediction with oracle-type theoretical guarantees.
Clustered Factor Analysis for Multivariate Spatial Data
Yanxiu Jin, Tomoya Wakayama, Renhe Jiang, Shonosuke Sugasawa
Spatial Statistics, 66, 100889 (2025)
A clustered factor analysis method for heterogeneous multivariate spatial data.
Spatiotemporal Factor Models for Functional Data with Application to Population Map Forecast
Tomoya Wakayama, Shonosuke Sugasawa
Spatial Statistics, 62, 100849 (2024)
A Bayesian spatio-temporal factor model for functional data with an application to population flow forecasting.
Functional Horseshoe Smoothing for Functional Trend Estimation
Tomoya Wakayama, Shonosuke Sugasawa
Statistica Sinica, 34(3), 1585–1602 (2024)
A shrinkage-prior approach to locally adaptive smoothing and trend estimation for functional time series.
Fast Convergence on Perfect Classification for Functional Data
Tomoya Wakayama, Masaaki Imaizumi
Statistica Sinica, 34(3), 1801–1819 (2024)
A theoretical analysis of perfect classification and exponential convergence for functional data classifiers.
Trend Filtering for Functional Data
Tomoya Wakayama, Shonosuke Sugasawa
Stat, 12(1), e590 (2023)
A trend filtering method for functional data based on ADMM, fused lasso, grouped lasso, and basis expansion.
Preprints
Open-ended Scientific Discovery with Possibilistic Reasoning
Anita Yang, Siu Lun Chau, Tomoya Wakayama, Krikamol Muandet, Masaki Adachi
Preprint, 2026
A possibility-theoretic framework for open-ended scientific discovery, introducing abductive utility and possibility frontier search with anytime-valid guarantees.
Diffusion Transformers are Provably Optimal In-context Generators
Guoji Fu, Tomoya Wakayama, Ryotaro Kawata, Atsushi Nitanda, Wee Sun Lee, Taiji Suzuki
Preprint, 2026
A theory of in-context generation showing that pretrained diffusion transformers can learn predictive distributions that capture task uncertainty and achieve minimax optimal rates.
The Geometry of Statistical Feature Learning in Mean-Field Langevin Dynamics
Zong Shang, Tomoya Wakayama, Guillaume Lecué, Taiji Suzuki
Preprint, 2026
A geometric formulation of statistical feature learning for supervised regression through mean-field Langevin dynamics.
On Misspecified Error Distributions in Bayesian Functional Clustering: Consequences and Remedies
Fumiya Iwashige, Tomoya Wakayama, Shonosuke Sugasawa, Shintaro Hashimoto
Preprint, 2025
A study of error misspecification in Bayesian functional clustering and remedies based on Gaussian process error modeling.
Process-based Inference for Spatial Energetics Using Bayesian Predictive Stacking
Tomoya Wakayama, Sudipto Banerjee
Preprint, 2024
Bayesian predictive stacking for process-based inference in spatial energetics and mobile health data.
Bayesian Inference for Consistent Predictions in Overparameterized Nonlinear Regression
Tomoya Wakayama
Preprint, 2024
Bayesian adaptive priors for consistent prediction and uncertainty quantification in overparameterized nonlinear regression.
Bayesian Analysis for Over-parameterized Linear Model via Effective Spectra
Tomoya Wakayama, Masaaki Imaizumi
Preprint, 2023
A Bayesian theory for overparameterized linear models without sparsity, using spectral priors and posterior contraction analysis.