Proceedings

Monte Carlo Filtering Objectives

Shuangshuang Chen, Sihao Ding, Yiannis Karayiannidis, Mårten Björkman

Download BibTeX Cite

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence. Main Track. Pages 2256-2262.

https://doi.org/10.24963/ijcai.2021/311

Keywords

  • Machine Learning: Learning Generative Models
  • Uncertainty in AI: Approximate Probabilistic Inference
  • Machine Learning: Unsupervised Learning
  • Machine Learning: Time-series; Data Streams

Abstract

Learning generative models and inferring latent trajectories have shown to be challenging for time series due to the intractable marginal likelihoods of flexible generative models. It can be addressed by surrogate objectives for optimization. We propose Monte Carlo filtering objectives (MCFOs), a family of variational objectives for jointly learning parametric generative models and amortized adaptive importance proposals of time series. MCFOs extend the choices of likelihood estimators beyond Sequential Monte Carlo in state-of-the-art objectives, possess important properties revealing the factors for the tightness of objectives, and allow for less biased and variant gradient estimates. We demonstrate that the proposed MCFOs and gradient estimations lead to efficient and stable model learning, and learned generative models well explain data and importance proposals are more sample efficient on various kinds of time series data.

Shuangshuang Chen, Sihao Ding, Yiannis Karayiannidis, and Mårten Björkman. 2021. Monte Carlo Filtering Objectives. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21. IJCAI, 2256–2262. https://doi.org/10.24963/ijcai.2021/311
Welcome Future Conferences AI Journal Division IJCAI Proceedings About IJCAI IJCAI Awards