DeepRV: Accelerating Spatiotemporal Inference with Pre-trained Neural Priors
arXiv:2503.21473v3 Announce Type: replace Abstract: Gaussian Processes (GPs) provide a flexible and statistically principled foundation for modelling spatiotemporal phenomena, but their $O(N^3)$ scaling makes them intractable for large datasets. Approximate methods such as variational inference (VI), inducing-point (sparse) GPs, low-rank kernel approximations (e.g., Nystrom methods and random Fourier features), and approximations such as INLA improve scalability but typically trade off accuracy, calibration, or modelling flexibility. We introduce DeepRV, a neural-network surrogate that replaces GP prior sampling, while closely matching full GP accuracy at inference including hyperparameter estimates, and reducing computational complexity to $O(N^2)$, increasing scalability and inference sp
View PDF HTML (experimental)
Abstract:Gaussian Processes (GPs) provide a flexible and statistically principled foundation for modelling spatiotemporal phenomena, but their $O(N^3)$ scaling makes them intractable for large datasets. Approximate methods such as variational inference (VI), inducing-point (sparse) GPs, low-rank kernel approximations (e.g., Nystrom methods and random Fourier features), and approximations such as INLA improve scalability but typically trade off accuracy, calibration, or modelling flexibility. We introduce DeepRV, a neural-network surrogate that replaces GP prior sampling, while closely matching full GP accuracy at inference including hyperparameter estimates, and reducing computational complexity to $O(N^2)$, increasing scalability and inference speed. DeepRV serves as a drop-in replacement for GP prior realisations in e.g. MCMC-based probabilistic programming pipelines, preserving full model flexibility. Across simulated benchmarks, non-separable spatiotemporal GPs, and a real-world application to education deprivation in London (n = 4,994 locations), DeepRV achieves the highest fidelity to exact GPs while substantially accelerating inference. Code is provided in the dl4bi Python package, with all experiments run on a single consumer-grade GPU to ensure accessibility for practitioners.
Comments: Code to reproduce all experiments is available in the dl4bi codebase: this https URL
Subjects:
Machine Learning (stat.ML); Machine Learning (cs.LG)
Cite as: arXiv:2503.21473 [stat.ML]
(or arXiv:2503.21473v3 [stat.ML] for this version)
https://doi.org/10.48550/arXiv.2503.21473
arXiv-issued DOI via DataCite
Submission history
From: Jhonathan Navott [view email] [v1] Thu, 27 Mar 2025 13:04:41 UTC (47,469 KB) [v2] Fri, 17 Oct 2025 16:20:58 UTC (11,913 KB) [v3] Tue, 31 Mar 2026 11:47:10 UTC (4,466 KB)
Sign in to highlight and annotate this article
Conversation starters
Daily AI Digest
Get the top 5 AI stories delivered to your inbox every morning.
More about
modelbenchmarkannounce
Q1 2026 Timelines Update
We’re mostly focused on research and writing for our next big scenario, but we’re also continuing to think about AI timelines and takeoff speeds, monitoring the evidence as it comes in, and adjusting our expectations accordingly. We’re tentatively planning on making quarterly updates to our timelines and takeoff forecasts. Since we published the AI Futures Model 3 months ago, we’ve updated towards shorter timelines. Daniel’s Automated Coder (AC) median has moved from late 2029 to mid 2028, and Eli’s forecast has moved a similar amount. The AC milestone is the point at which an AGI company would rather lay off all of their human software engineers than stop using AIs for software engineering. The reasons behind this change include: 1 We switched to METR Time Horizon version 1.1 . We include
Knowledge Map
Connected Articles — Knowledge Graph
This article is connected to other articles through shared AI topics and tags.
More in Products
2026: The year of throwing my agency at my health (now with added cyborgism)
I have bipolar disorder. I was diagnosed in late 2012 following my one and only severe manic episode. Most psychiatrists would regard me as a resounding success case – I never even remotely come close to suicidal depression, manic delusions of grandeur, impulsive spending, or irresponsible sexual behavior. By standard measures, I am well-adjusted, functional, and successful. Part of this relative success is adherence to appropriate medication, and another part is maintaining good insight [1] into my mental state. Years ago, I defined a personal bipolar index scale to communicate to myself and close ones my mental state. My bipolar index ranges from -10 to +10 and is a subjective self-report. -10 would be a state of extreme suicidal depression. +10 would be extreme mania with complete loss
Microsoft execs warn agentic AI is hollowing out the junior developer pipeline
Two of Microsoft s most prominent developers have some words for organizations overly celebrating agentic AI s productivity gains: You are hollowing The post Microsoft execs warn agentic AI is hollowing out the junior developer pipeline appeared first on The New Stack .
Discussion
Sign in to join the discussion
No comments yet — be the first to share your thoughts!