Likelihood-Free Inference in State-Space Models with Unknown Dynamics

Alexander Aushev, Thong Tran, Henri Pesonen, Andrew Howes, Samuel Kaski

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


We introduce a method for inferring and predicting latent states in the important and difficult case of state-space models where observations can only be simulated, and transition dynamics are unknown. In this setting, the likelihood of observations is not available and only synthetic observations can be generated from a black-box simulator. We propose a way of doing likelihood-free inference (LFI) of states and state prediction with a limited number of simulations. Our approach uses a multi-output Gaussian process for state inference, and a Bayesian Neural Network as a model of the transition dynamics for state prediction. We improve upon existing LFI methods for the inference task, while also accurately learning transition dynamics. The proposed method is necessary for modelling inverse problems in dynamical systems with computationally expensive simulations, as demonstrated in experiments with non-stationary user models.
Original languageEnglish
Title of host publicationpre-print
Number of pages23
Publication statusPublished - 20 Feb 2023
EventNeurIPS 2021 Workshop on Deep Learning and Inverse Problems - Virtual-only Conference, United States
Duration: 6 Dec 202114 Dec 2021


ConferenceNeurIPS 2021 Workshop on Deep Learning and Inverse Problems
Country/TerritoryUnited States
Internet address

Research Beacons, Institutes and Platforms

  • Institute for Data Science and AI
  • Digital Futures
  • Christabel Pankhurst Institute


Dive into the research topics of 'Likelihood-Free Inference in State-Space Models with Unknown Dynamics'. Together they form a unique fingerprint.

Cite this