---
title: "Gibbs Variational Last Layers for Long-Horizon Dynamics Learning"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2611.02891v1"
permalink: "https://feed.astralor.com/items/itm_180afcc920f3d6f1af62f723"
published: "2026-11-02T10:25:14.000Z"
score: 57
channel: "tech"
tags: ["推理效率","Gibbs Variational Last","Layers","Long-Horizon Dynamics Learning","Accurate","Variational Bayesian","VBLLs"]
---

# Gibbs Variational Last Layers for Long-Horizon Dynamics Learning

## 速览

Accurate long-horizon probabilistic predictions of dynamical systems are a prerequisite for model-based decision-making. Variational Bayesian last layers (VBLLs) are an attractive model class for this setting, offering …

## 关键点

- Accurate long-horizon probabilistic predictions of dynamical systems are a prerequisite for model-based decision-making.
- Variational Bayesian last layers (VBLLs) are an attractive model class for this setting, offering tractable uncertainty at near-deterministic cost, yet they are trained exclusively with single-step likelihood objectives…
- We revisit VBLLs through the lens of Gibbs variational inference, which decouples posterior inference from the choice of training loss.

## 全文

Accurate long-horizon probabilistic predictions of dynamical systems are a prerequisite for model-based decision-making. Variational Bayesian last layers (VBLLs) are an attractive model class for this setting, offering tractable uncertainty at near-deterministic cost, yet they are trained exclusively with single-step likelihood objectives that may be insufficient for reliable multi-step rollouts. We revisit VBLLs through the lens of Gibbs variational inference, which decouples posterior inference from the choice of training loss. Building on this perspective, we introduce a family of multi-step training losses that progressively incorporate long-horizon structure, robustness via continuous ranked probability scores (CRPS), and training on full autoregressive rollouts. Our approach retains the simplicity of VBLLs while directly optimizing for multi-step predictive quality. Across illustrative synthetic environments, a diverse suite of chaotic dynamical systems, and real-world data, we show that multi-step, CRPS-based objectives substantially improve long-horizon accuracy and distributional fit, consistently outperforming single-step likelihood training.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_180afcc920f3d6f1af62f723