ADELIA: Automatic Differentiation for Efficient Laplace Inference Approximations
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913098959421440 |
|---|---|
| author | Boudaoud, Afif Gaedke-Merzhäuser, Lisa Ziogas, Alexandros Nikolaos Maillou, Vincent Calotoiu, Alexandru Copik, Marcin Rue, Håvard Luisier, Mathieu Hoefler, Torsten |
| author_facet | Boudaoud, Afif Gaedke-Merzhäuser, Lisa Ziogas, Alexandros Nikolaos Maillou, Vincent Calotoiu, Alexandru Copik, Marcin Rue, Håvard Luisier, Mathieu Hoefler, Torsten |
| contents | Spatio-temporal Bayesian inference drives environmental and health sciences using latent Gaussian models. Integrated Nested Laplace Approximations (INLA) enable inference for these models at HPC scale but rely on derivative-based optimization over $d$ hyperparameters. State-of-the-art INLA implementations approximate derivatives via central finite differences (FD), requiring $2d{+}1$ evaluations. These evaluations are embarrassingly parallel, but total work and energy grow with $d$, limiting time-to-solution under fixed budgets. Reverse-mode automatic differentiation (AD) computes exact gradients independently of $d$, but its efficient application to INLA's structured-sparse kernels is an open challenge. We present ADELIA, the first AD-enabled INLA implementation with a structure-exploiting multi-GPU backward pass leveraging model sparsity. We evaluate ADELIA on ten benchmark models, including real-world air-pollution monitoring. We achieve $4.2$--$7.9\times$ per-gradient speedups and reliable convergence on production-scale models with up to 1.9M latent variables, where FD struggles. Even when scaled to 16--32 GPUs to match ADELIA's wall-clock time, FD consumes $5$--$8\times$ more energy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_06392 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ADELIA: Automatic Differentiation for Efficient Laplace Inference Approximations Boudaoud, Afif Gaedke-Merzhäuser, Lisa Ziogas, Alexandros Nikolaos Maillou, Vincent Calotoiu, Alexandru Copik, Marcin Rue, Håvard Luisier, Mathieu Hoefler, Torsten Distributed, Parallel, and Cluster Computing Performance Spatio-temporal Bayesian inference drives environmental and health sciences using latent Gaussian models. Integrated Nested Laplace Approximations (INLA) enable inference for these models at HPC scale but rely on derivative-based optimization over $d$ hyperparameters. State-of-the-art INLA implementations approximate derivatives via central finite differences (FD), requiring $2d{+}1$ evaluations. These evaluations are embarrassingly parallel, but total work and energy grow with $d$, limiting time-to-solution under fixed budgets. Reverse-mode automatic differentiation (AD) computes exact gradients independently of $d$, but its efficient application to INLA's structured-sparse kernels is an open challenge. We present ADELIA, the first AD-enabled INLA implementation with a structure-exploiting multi-GPU backward pass leveraging model sparsity. We evaluate ADELIA on ten benchmark models, including real-world air-pollution monitoring. We achieve $4.2$--$7.9\times$ per-gradient speedups and reliable convergence on production-scale models with up to 1.9M latent variables, where FD struggles. Even when scaled to 16--32 GPUs to match ADELIA's wall-clock time, FD consumes $5$--$8\times$ more energy. |
| title | ADELIA: Automatic Differentiation for Efficient Laplace Inference Approximations |
| topic | Distributed, Parallel, and Cluster Computing Performance |
| url | https://arxiv.org/abs/2605.06392 |