Learning High-Dimensional Mixed Models via Amortized Variational Inference

Abstract

Modelling longitudinal data is an important yet challenging task. These datasets can be high-dimensional, consist of non-linear effects, and contain time-varying covariates. In this work, we leverage linear mixed models (LMMs) and amortized variational inference to provide conditional priors for VAEs, and propose LMM-VAE, a model that is scalable, interpretable, and shares theoretical connections to the GP-based VAEs. We empirically demonstrate that LMM-VAE performs competitively compared to existing approaches.

Cite

Text

Ong et al. "Learning High-Dimensional Mixed Models via Amortized Variational Inference." ICML 2024 Workshops: SPIGM, 2024.

Markdown

[Ong et al. "Learning High-Dimensional Mixed Models via Amortized Variational Inference." ICML 2024 Workshops: SPIGM, 2024.](https://mlanthology.org/icmlw/2024/ong2024icmlw-learning/)

BibTeX

@inproceedings{ong2024icmlw-learning,
  title     = {{Learning High-Dimensional Mixed Models via Amortized Variational Inference}},
  author    = {Ong, Priscilla and Haussmann, Manuel and Lähdesmäki, Harri},
  booktitle = {ICML 2024 Workshops: SPIGM},
  year      = {2024},
  url       = {https://mlanthology.org/icmlw/2024/ong2024icmlw-learning/}
}