Train Stochastic Non Linear Coupled ODEs to Classify and Generate
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915553559445504 |
|---|---|
| author | Gagliani, Stefano Pacifico, Feliciano Giuseppe Chicchi, Lorenzo Fanelli, Duccio Febbe, Diego Buffoni, Lorenzo Marino, Raffaele |
| author_facet | Gagliani, Stefano Pacifico, Feliciano Giuseppe Chicchi, Lorenzo Fanelli, Duccio Febbe, Diego Buffoni, Lorenzo Marino, Raffaele |
| contents | A general class of dynamical systems which can be trained to operate in classification and generation modes are introduced. A procedure is proposed to plant asymptotic stationary attractors of the deterministic model. Optimizing the dynamical system amounts to shaping the architecture of inter-nodes connection to steer the evolution towards the assigned equilibrium, as a function of the class to which the item - supplied as an initial condition - belongs to. Under the stochastic perspective, point attractors are turned into probability distributions, made analytically accessible via the linear noise approximation. The addition of noise proves beneficial to oppose adversarial attacks, a property that gets engraved into the trained adjacency matrix and therefore also inherited by the deterministic counterpart of the optimized stochastic model. By providing samples from the target distribution as an input to a feedforward neural network (or even to a dynamical model of the same typology of the adopted for classification purposes), yields a fully generative scheme. Conditional generation is also possible by merging classification and generation modalities. Automatic disentanglement of isolated key features is finally proven. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_12286 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Train Stochastic Non Linear Coupled ODEs to Classify and Generate Gagliani, Stefano Pacifico, Feliciano Giuseppe Chicchi, Lorenzo Fanelli, Duccio Febbe, Diego Buffoni, Lorenzo Marino, Raffaele Disordered Systems and Neural Networks Statistical Mechanics A general class of dynamical systems which can be trained to operate in classification and generation modes are introduced. A procedure is proposed to plant asymptotic stationary attractors of the deterministic model. Optimizing the dynamical system amounts to shaping the architecture of inter-nodes connection to steer the evolution towards the assigned equilibrium, as a function of the class to which the item - supplied as an initial condition - belongs to. Under the stochastic perspective, point attractors are turned into probability distributions, made analytically accessible via the linear noise approximation. The addition of noise proves beneficial to oppose adversarial attacks, a property that gets engraved into the trained adjacency matrix and therefore also inherited by the deterministic counterpart of the optimized stochastic model. By providing samples from the target distribution as an input to a feedforward neural network (or even to a dynamical model of the same typology of the adopted for classification purposes), yields a fully generative scheme. Conditional generation is also possible by merging classification and generation modalities. Automatic disentanglement of isolated key features is finally proven. |
| title | Train Stochastic Non Linear Coupled ODEs to Classify and Generate |
| topic | Disordered Systems and Neural Networks Statistical Mechanics |
| url | https://arxiv.org/abs/2510.12286 |