Latent jump models enable faster learning of event-driven time series
Scan-Parallel Variational Inference for event-driven Latent Jump Models
Machine Learning
Summary
Time series data often has irregular measurements and sudden changes that are hard to capture with smooth models. The authors introduce latent jump models where hidden states change instantly at random event times, better representing sudden shocks. They developed a new way to efficiently train these models that speeds up learning by looking at events in parallel instead of step-by-step. Their approach works well on synthetic and real irregular time data, often faster than traditional methods without losing accuracy.
What this means in practice
- •For time series data engineers: Train models that capture sudden events in irregular time series faster than traditional continuous-time approaches while keeping forecast accuracy.
- •For financial risk analysts: Identify and predict the timing and impact of rare but significant market shocks using more efficient latent jump modeling techniques.
Authors
Vansh Bansal, Min Chen, James G. Scott
Abstract
Continuous-time latent state-space models are widely used for irregularly observed time series, typically representing uncertainty through a random initial state followed by smooth ordinary differential equation (ODE) dynamics, or Brownian perturbations accumulated by a stochastic differential equation (SDE). While these constructions are natural for smooth or diffusive systems, they do not directly encode a finite set of unobserved shocks whose timing, magnitude, and effect on memory are themselves uncertain. We instead introduce latent jump models (LJMs): the latent state follows closed-form deterministic dynamics between Poisson-driven event times and undergoes a learned instantaneous jump at each event, with state-dependent intensity and mark distributions. This defines a time-homogeneous Markov process whose jumps are intrinsic state transitions rather than numerical integration steps. For training, we derive a tractable event-augmented path-space evidence lower bound (ELBO) and introduce Scan-Parallel Variational Inference (SPVI): an observation-conditioned variational posterior samples the complete latent event configuration before the state recurrence is evaluated using triangular affine scans. This enables fast and scalable training for LJMs in logarithmic temporal depth. Across synthetic stochastic systems and real irregular time series, LJMs remain competitive with continuous-time latent baselines in forecast accuracy, often with fewer latent updates than numerical solver evaluations, while training one to two orders of magnitude faster than the sequential baselines.