Overview

Much real-world data such as medical records, customer interactions, or financial transactions is recorded at irregular time intervals.  However, most deep learning time series models, such as recurrent neural networks, require data to be recorded at regular intervals, such hourly or daily.  I'll explain some recent advances in building deep stochastic differential equation models that specify continuous-time dynamics.  This allows us to fit a new family of richly-parameterized distributions over time series.  We'll discuss their strengths and limitations, and demonstrate these models on medical records and motion capture data.  All the tools discussed are open-source.

Session Overview

  • 01

    ODSC West 2020: Continuous-time Deep Models for Forecasting Sparse Time Series

    • Abstract & Bio

    • Continuous-time Deep Models for Forecasting Sparse Time Series

INTERESTED IN HANDS-ON TRAINING SESSIONS?

Start your 14-days trial. Cancel anytime.