7 papers
A Markov Chain Approach to Preference Alignment
Takuya Koriyama, Tengyuan Liang
We propose Markov Chain from Human Feedback (MCHF), an elementary approach for aligning generative models from pairwise human preferences. Unlike Reinforcement Learning from Human…
Denoising Diffusions with Optimal Transport: Localization, Curvature, and Multi-Scale Complexity
Tengyuan Liang, Kulunu Dharmakeerthi, Takuya Koriyama
Adding noise is easy; what about denoising? Diffusion is easy; what about reverting a diffusion? Diffusion-based generative models aim to denoise a Langevin diffusion chain, moving…
Asymptotics of resampling without replacement in robust and logistic regression
Pierre C. Bellec, Takuya Koriyama
This paper studies the asymptotics of resampling without replacement in the proportional regime where dimension and sample size are of the same order. For a given dataset $…
Precise Asymptotics of Bagging Regularized M-estimators
Takuya Koriyama, Pratik Patil, Jin-Hong Du +2
We characterize the squared prediction risk of ensemble estimators obtained through subagging (subsample bootstrap aggregating) regularized M-estimators and construct a consistent…
Phase transitions for the existence of unregularized M-estimators in single index models
Takuya Koriyama, Pierre C. Bellec
This paper studies phase transitions for the existence of unregularized M-estimators under proportional asymptotics where the sample size and feature dimension grow proport…
Error estimation and adaptive tuning for unregularized robust M-estimator
Pierre C. Bellec, Takuya Koriyama
We consider unregularized robust M-estimators for linear models under Gaussian design and heavy-tailed noise, in the proportional asymptotics regime where the sample size n and the…