Research map / Time series / Causal and generative time series

Koopman Generative Operators for Efficient Probabilistic Time-Series Forecasting

NeurIPS 2026 · Time series › Causal and generative time series · approach: Diffusion models for time series

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes Koopman-based generative operators for probabilistic time-series forecasting that combine structured latent dynamics with expressive, efficient generation.

Problem

Probabilistic time-series forecasting needs models that jointly capture structured temporal dynamics, expressive uncertainty, and efficient inference; latent dynamical models are restrictive in generation, while diffusion and flow matching are iterative and computationally costly.

Approach

Uses Koopman operator theory to model time-series dynamics in a latent space with a generative operator, aiming to retain structured dynamics while avoiding iterative sampling of diffusion/flow-matching models. (Details inferred from abstract only.)

Experimental setup

Baselines: diffusion models flow matching

Closest papers by meaning

• $K^2$VAE: A Koopman-Kalman Enhanced Variational AutoEncoder for Probabilistic Time Series Forecasting (ICML 2025 · similarity 0.91)
• Probabilistic Forecasting with Stochastic Interpolants and Föllmer Processes (ICML 2024 · similarity 0.89)
• SKOLR: Structured Koopman Operator Linear RNN for Time-Series Forecasting (ICML 2025 · similarity 0.88)
• Generative Modeling of Regular and Irregular Time Series Data via Koopman VAEs (ICLR 2024 · similarity 0.88)
• Flow Matching with Gaussian Process Priors for Probabilistic Time Series Forecasting (ICLR 2025 · similarity 0.88)
• KoNODE: Koopman-Driven Neural Ordinary Differential Equations with Evolving Parameters for Time Series Analysis (ICML 2025 · similarity 0.88)