arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.14833cs.LGcs.SYeess.SY
收藏

深度结构化状态空间模型的模型降阶:一种系统理论方法

Model order reduction of deep structured state-space models: A system-theoretic approach

IDSIA Dalle Molle Institute for Artificial Intelligence USI-SUPSI(IDSIA 达尔·莫勒人工智能研究所(USI-SUPSI))

Marco Forgione, Manas Mejari, Dario Piga

AI总结:

针对深度结构化状态空间模型因阶数过大而不适合控制设计的问题,提出基于系统理论的模型降阶方法,通过引入模态 $\ell_1$ 和 Hankel 核范数正则化促进稀疏性,在保持精度的同时实现精简表示与更快推理。

摘要:

在参数化系统辨识中,以控制设计目标为特定重点,在有限复杂度下实现精确的系统建模至关重要。近期引入的深度结构化状态空间模型(SSM)以线性动态块作为关键组成成分,提供了高预测性能。然而,学习到的表示通常具有过大的模型阶数,使其不适合控制设计目的。本文通过针对SSM线性动态块的系统理论模型降阶技术来解决这一挑战。我们引入了两个正则化项,可将其纳入训练损失以改进模型降阶。具体而言,我们考虑模态 $\ell_1$ 和 Hankel 核范数正则化以促进稀疏性,从而仅保留相关状态而不牺牲精度。所提出的正则化器在精简表示和由降阶模型带来的更快推理方面具有优势。使用来自飞机的真实地面振动数据证明了所提方法的有效性。

英文摘要:

With a specific emphasis on control design objectives, achieving accurate system modeling with limited complexity is crucial in parametric system identification. The recently introduced deep structured state-space models (SSM), which feature linear dynamical blocks as key constituent components, offer high predictive performance. However, the learned representations often suffer from excessively large model orders, which render them unsuitable for control design purposes. The current paper addresses this challenge by means of system-theoretic model order reduction techniques that target the linear dynamical blocks of SSMs. We introduce two regularization terms which can be incorporated into the training loss for improved model order reduction. In particular, we consider modal $\ell_1$ and Hankel nuclear norm regularization to promote sparsity, allowing one to retain only the relevant states without sacrificing accuracy. The presented regularizers lead to advantages in terms of parsimonious representations and faster inference resulting from the reduced order models. The effectiveness of the proposed methodology is demonstrated using real-world ground vibration data from an aircraft.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑