AI & ML
Generative sequence modeling for infinite memory processes via predictive states
arXiv stat.ML0 views
arXiv:2609.38524v3 Announce Type: replace Abstract: We consider estimating the one-step-ahead conditional distribution of a multivariate stochastic process. Many existing approaches rely on assumptions such as finite-range memory, sparsity, or additivity, which can be poorly suited to processes with long-range nonlinear interactions. However, without such structural assumptions, nonparametric estimation is challenging due to the curse of dimensionality. To address this challenge, we introduce a new estimator based on learning representations of the \emph{predictive states} of a stochastic process. We show that the statistical complexity of the estimation problem is determined by the dimension of this predictive state representation. Furthermore, we show that a $p$-dimensional predictive state space can be represented in a $2p+1$-dimensional feature space. Our results imply that faster convergence rates are achievable when the predictive state space is lower-dimensional, i.e., when the past history can be compressed into a lower-dimensional statistic that is sufficient for predicting the future. We establish theoretical convergence rate guarantees for an instantiation of our method based on deep neural network estimators. We implement our method in PyTorch and evaluate its performance on synthetic data and real \emph{C. elegans} data.
Source:https://arxiv.org/abs/2609.38524Previousmlr3torch: A Deep Learning Framework in R based on mlr3 and torchNextTrajectory Density Ratios under Weak Support: A Survey across the Sciences
Related posts
One Coding Plan, Two CLIs: Configure Claude Code and Codex Without Credential Collisions
0dev.to Programming
200 countries: Google Maps just became an agent
0dev.to AI
MiniMax H3 Long Video: The Human-in-the-Loop Editing Pattern
0dev.to AI
From Artificial Neurons to AlexNet and Beyond.
0dev.to AI

국내 대표 AI 기업 11개사, 2026년 허깅페이스 공개 모델 다운로드 분석
0dev.to AI
Token Cost Reality Check: Running Xinference for Multi-Model RAG Without Breaking the Budget
0dev.to Python