# Deep Learning of Dynamic Systems using System Identification Toolbox(TM)

arXiv: https://arxiv.org/abs/2409.07642 · HTML: https://mlsanity.com/paper/2409.07642/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 13 references that are in the map.
- First posted on arXiv 11 Sep 2024.

## Builds on (most cited first, 3 of 3)

- [Neural Ordinary Differential Equations](https://mlsanity.com/paper/1806.07366/) (2018)
- [Deep networks for system identification: a Survey](https://mlsanity.com/paper/2301.12832/) (2023)
- [Gradient Hard Thresholding Pursuit for Sparsity-Constrained Optimization](https://mlsanity.com/paper/1311.5750/) (2013)

## Built on by (most cited first, 5 of 5)

- [Stable-by-Design Neural Network-Based LPV State-Space Models for System Identification](https://mlsanity.com/paper/2510.24757/) (2025)
- [SOLIS: Physics-Informed Learning of Interpretable Neural Surrogates for Nonlinear Systems](https://mlsanity.com/paper/2604.14879/) (2026)
- [xFODE+: Explainable Type-2 Fuzzy Additive ODEs for Uncertainty Quantification](https://mlsanity.com/paper/2604.14880/) (2026)
- [xFODE: An Explainable Fuzzy Additive ODE Framework for System Identification](https://mlsanity.com/paper/2604.14883/) (2026)
- [Beyond Prediction: Interval Neural Networks for Uncertainty-Aware System Identification](https://mlsanity.com/paper/2605.11460/) (2026)

## Abstract (arXiv)

MATLAB(R) releases over the last 3 years have witnessed a continuing growth in the dynamic modeling capabilities offered by the System Identification Toolbox(TM). The emphasis has been on integrating deep learning architectures and training techniques that facilitate the use of deep neural networks as building blocks of nonlinear models. The toolbox offers neural state-space models which can be extended with auto-encoding features that are particularly suited for reduced-order modeling of large systems. The toolbox contains several other enhancements that deepen its integration with the state-of-art machine learning techniques, leverage auto-differentiation features for state estimation, and enable a direct use of raw numeric matrices and timetables for training models.