PulseAugur
EN
LIVE 09:15:46

New generative models tackle scarce time series data and predictive distributions

Two new research papers propose advanced generative models for time series data. The first, TimeMoDE, utilizes Diffusion Transformers with Mixture-of-Experts to effectively generate time series even with scarce data by leveraging domain knowledge and diffusion stage awareness. The second paper introduces a framework for modeling predictive distributions of time series using conditional generative adversarial networks, enabling robust forecasting and risk assessment with efficient computation. AI

IMPACT These papers advance generative modeling techniques for time series, potentially improving forecasting accuracy and data synthesis capabilities in data-scarce environments.

RANK_REASON Two distinct arXiv papers introducing new methodologies for time series generation and prediction.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 3 sources. How we write summaries →

COVERAGE [3]

  1. arXiv cs.LG TIER_1 English(EN) · Zihao Yao, Qi Zheng, Jiankai Zuo, Yaying Zhang ·

    Towards a Unified Generative Model for Scarce Time Series with Domain Experts

    arXiv:2606.15172v1 Announce Type: new Abstract: Synthesizing realistic time series with generative models has wide-ranging applications in real-world scenarios. Despite recent progress, most existing methods are trained under the assumption of abundant training data, which substa…

  2. arXiv stat.ML TIER_1 English(EN) · Jordi Llorens-Terrazas, Mika Meitz ·

    Generative Predictive Distributions for Time Series

    arXiv:2606.16773v1 Announce Type: cross Abstract: We propose a flexible framework for modeling the predictive distributions of nonlinear, possibly multivariate time series. Our approach expresses a general predictive distribution in an appropriate generative representation that i…

  3. arXiv stat.ML TIER_1 English(EN) · Mika Meitz ·

    Generative Predictive Distributions for Time Series

    We propose a flexible framework for modeling the predictive distributions of nonlinear, possibly multivariate time series. Our approach expresses a general predictive distribution in an appropriate generative representation that is based on a folklore result from measure theoreti…