Menu
Home Explore People Places Arts History Plants & Animals Science Life & Culture Technology
On this page
Neural differential equation
Equation in machine learning

In machine learning, a neural differential equation is a differential equation whose right-hand side is parametrized by the weights θ of a neural network. In particular, a neural ordinary differential equation (neural ODE) is an ordinary differential equation of the form

d h ( t ) d t = f θ ( h ( t ) , t ) . {\displaystyle {\frac {\mathrm {d} \mathbf {h} (t)}{\mathrm {d} t}}=f_{\theta }(\mathbf {h} (t),t).} In classical neural networks, layers are arranged in a sequence indexed by natural numbers. In neural ODEs, however, layers form a continuous family indexed by positive real numbers. Specifically, the function h : R ≥ 0 → R {\displaystyle h:\mathbb {R} _{\geq 0}\to \mathbb {R} } maps each positive index t to a real value, representing the state of the neural network at that layer.

Neural ODEs can be understood as continuous-time control systems, where their ability to interpolate data can be interpreted in terms of controllability.

We don't have any images related to Neural differential equation yet.
We don't have any YouTube videos related to Neural differential equation yet.
We don't have any PDF documents related to Neural differential equation yet.
We don't have any Books related to Neural differential equation yet.
We don't have any archived web articles related to Neural differential equation yet.

Connection with residual neural networks

Neural ODEs can be interpreted as a residual neural network with a continuum of layers rather than a discrete number of layers.3 Applying the Euler method with a unit time step to a neural ODE yields the forward propagation equation of a residual neural network:

h ℓ + 1 = f θ ( h ℓ , ℓ ) + h ℓ , {\displaystyle \mathbf {h} _{\ell +1}=f_{\theta }(\mathbf {h} _{\ell },\ell )+\mathbf {h} _{\ell },}

with ℓ being the ℓ-th layer of this residual neural network. While the forward propagation of a residual neural network is done by applying a sequence of transformations starting at the input layer, the forward propagation computation of a neural ODE is done by solving a differential equation. More precisely, the output h out {\displaystyle \mathbf {h} _{\text{out}}} associated to the input h in {\displaystyle \mathbf {h} _{\text{in}}} of the neural ODE is obtained by solving the initial value problem

d h ( t ) d t = f θ ( h ( t ) , t ) , h ( 0 ) = h in , {\displaystyle {\frac {\mathrm {d} \mathbf {h} (t)}{\mathrm {d} t}}=f_{\theta }(\mathbf {h} (t),t),\quad \mathbf {h} (0)=\mathbf {h} _{\text{in}},}

and assigning the value h ( T ) {\displaystyle \mathbf {h} (T)} to h out {\displaystyle \mathbf {h} _{\text{out}}} .

Universal differential equations

In physics-informed contexts where additional information is known, neural ODEs can be combined with an existing first-principles model to build a physics-informed neural network model called universal differential equations (UDE).4567 For instance, an UDE version of the Lotka-Volterra model can be written as8

d x d t = α x − β x y + f θ ( x ( t ) , y ( t ) ) , d y d t = − γ y + δ x y + g θ ( x ( t ) , y ( t ) ) , {\displaystyle {\begin{aligned}{\frac {dx}{dt}}&=\alpha x-\beta xy+f_{\theta }(x(t),y(t)),\\{\frac {dy}{dt}}&=-\gamma y+\delta xy+g_{\theta }(x(t),y(t)),\end{aligned}}}

where the terms f θ {\displaystyle f_{\theta }} and g θ {\displaystyle g_{\theta }} are correction terms parametrized by neural networks.

See also

References

  1. Chen, Ricky T. Q.; Rubanova, Yulia; Bettencourt, Jesse; Duvenaud, David K. (2018). "Neural Ordinary Differential Equations" (PDF). In Bengio, S.; Wallach, H.; Larochelle, H.; Grauman, K.; Cesa-Bianchi, N.; Garnett, R. (eds.). Advances in Neural Information Processing Systems. Vol. 31. Curran Associates, Inc. arXiv:1806.07366. https://proceedings.neurips.cc/paper_files/paper/2018/file/69386f6bb1dfed68692a24c8686939b9-Paper.pdf

  2. Ruiz-Balet, Domènec; Zuazua, Enrique (2023). "Neural ODE Control for Classification, Approximation, and Transport". SIAM Review. 65 (3): 735–773. arXiv:2104.05278. doi:10.1137/21M1411433. ISSN 0036-1445. https://epubs.siam.org/doi/10.1137/21M1411433

  3. Chen, Ricky T. Q.; Rubanova, Yulia; Bettencourt, Jesse; Duvenaud, David K. (2018). "Neural Ordinary Differential Equations" (PDF). In Bengio, S.; Wallach, H.; Larochelle, H.; Grauman, K.; Cesa-Bianchi, N.; Garnett, R. (eds.). Advances in Neural Information Processing Systems. Vol. 31. Curran Associates, Inc. arXiv:1806.07366. https://proceedings.neurips.cc/paper_files/paper/2018/file/69386f6bb1dfed68692a24c8686939b9-Paper.pdf

  4. Christopher Rackauckas; Yingbo Ma; Julius Martensen; Collin Warner; Kirill Zubov; Rohit Supekar; Dominic Skinner; Ali Ramadhan; Alan Edelman (2024). "Universal Differential Equations for Scientific Machine Learning". arXiv:2001.04385 [cs.LG]. /wiki/ArXiv_(identifier)

  5. Xiao, Tianbai; Frank, Martin (2023). "RelaxNet: A structure-preserving neural network to approximate the Boltzmann collision operator". Journal of Computational Physics. 490: 112317. arXiv:2211.08149. Bibcode:2023JCoPh.49012317X. doi:10.1016/j.jcp.2023.112317. https://linkinghub.elsevier.com/retrieve/pii/S0021999123004126

  6. Silvestri, Mattia; Baldo, Federico; Misino, Eleonora; Lombardi, Michele (2023), Mikyška, Jiří; de Mulatier, Clélia; Paszynski, Maciej; Krzhizhanovskaya, Valeria V. (eds.), "An Analysis of Universal Differential Equations for Data-Driven Discovery of Ordinary Differential Equations", Computational Science – ICCS 2023, vol. 10476, Cham: Springer Nature Switzerland, pp. 353–366, doi:10.1007/978-3-031-36027-5_27, ISBN 978-3-031-36026-8, retrieved 2024-08-18 978-3-031-36026-8

  7. Christoph Plate; Carl Julius Martensen; Sebastian Sager (2024). "Optimal Experimental Design for Universal Differential Equations". arXiv:2408.07143 [math.OC]. /wiki/ArXiv_(identifier)

  8. Patrick Kidger (2021). On Neural Differential Equations (PhD). Oxford, United Kingdom: University of Oxford, Mathematical Institute. https://ora.ox.ac.uk/objects/uuid:af32d844-df84-4fdc-824d-44bebc3d7aa9