DS3M: 面向现代硬件的非线性系统辨识深度结构化状态空间模型
DS3M: Deep structured state space models for nonlinear system identification on modern hardware
Marco Roschkowski · Hannes Gernandt
中文摘要
神经状态空间辨识从测量的输入输出中学习动力学模型。SUBNET 将这一任务与编码器结合,由编码器从短测量历史估计初始状态。其连续时间版本在数值积分过程中需要大量连续的神经网络评估,导致训练速度缓慢,这是实际应用中的主要限制。提出深度结构化状态空间模型(DS3M),将紧凑的非线性动力学组件以并联和串联方式组合。使用 Triton 实现的自定义 GPU 内核降低了其顺序评估的启动和内存开销。阐述该构造,描述一种 port-Hamiltonian 变体,并给出多项关于精度和训练时间的基准比较。
关键要点
- 01SUBNET 连续时间版本因数值积分需大量网络评估,训练速度慢
- 02DS3M 将紧凑非线性动力学组件并联与串联组合
- 03用 Triton 编写自定义 GPU 内核,降低顺序评估的启动与内存开销
- 04给出 port-Hamiltonian 变体及精度与训练时间的基准比较
解读
尚无解读。
原始英文摘要
arXiv:2611.25274v1 Announce Type: new Abstract: Neural state-space identification learns dynamical models from measured inputs and outputs. SUBNET combines this task with an encoder that estimates initial states from short measurement histories. Its continuous-time version requires many successive neural-network evaluations during numerical integration, resulting in slow training speed which is a major limitation for practical use cases. We propose deep structured state space models (DS3M), which combine compact nonlinear dynamical components in parallel and in series. Custom GPU kernels implemented with Triton reduce the launch and memory overhead of their sequential evaluation. We explain this construction, describe a port-Hamiltonian variant, and present various benchmark comparisons of accuracy and training time.