STAFF
HONORARY STAFF
POSTDOCS
PHD STUDENTS
VISITORS
RELATED STAFF
|
| Name |
Title |
Date |
Time |
Room |
| Eugenio Pozzoli (Université de Rennes) | Controllability of Hamiltonian diffeomorphisms and mechanical Liouville equationsAbstract: In this talk, we consider a time-dependent mechanical Hamiltonian (i.e. kinetic plus potential energy) on a cotangent bundle manifold, with a potential that can be modulated in time by the choice of a control function. The flow of the associated controlled Hamiltonian system is a path inside the group of Hamiltonian diffeomorphisms. We propose an example of mechanical Hamiltonian, on the cotangent bundle of the Euclidean space, for which any Hamiltonian diffeomorphism can be attained (in an approximate sense, and in arbitrarily small times) via the flow of the controlled system, by a suitable choice of control function.
We also study another example on the cotangent bundle of the torus. In this case, we only prove a weaker property, that is the small-time approximate controllability of the associated Liouville equation: any two densities (whose level sets have the same measure) can be connected via the controlled Liouville transport equation.
This talk is based on the preprint [1] in collaboration with Bettina Kazandjian and Mario Sigalotti (Inria and LJLL, Sorbonne Université, Paris), and is the classical analogue of a previous work [2] on Schrödinger equations in collaboration with Karine Beauchard (ENS Rennes).
[1] B. Kazandjian, E. Pozzoli, M. Sigalotti; Orbits and attainable Hamiltonian diffeomorphisms for mechanical Liouville equations. arXiv:2509.24960 (2025).
[2] K. Beauchard, E. Pozzoli; Small-time approximate controllability of Schrödinger equations and diffeomorphisms. Ann. Inst. H. Poincaré Analyse Non-Linéaire (2025). |
Wednesday, 5 November 2025 |
11:00 |
HXLY 410 |
| Zemer Kosloff (University of Bristol, Hebrew University of Jerusalem) | The ergodic index of non-stationary Bernoulli shiftsAbstract: A non-stationary Bernoulli shift is the shift action of an independent, not necessarily identically distributed random variables under the assumption that the push forward of the distribution (product measure) of the shift preserves the null-sets of the distribution.
These shifts were first considered by Krengel in 1969 and Hamachi (1981) who showed that there exists non-stationary Bernoulli shift examples which do not have an absolutely continuous shift invariant distribution.
We will survey some of the recent progress on understanding the (variety of possible) ergodic theoretic behaviour of this model and present a result which is joint with Terry Soo which shows that there exists a Bernoulli shift of an arbitrary ergodic index. On the way we expect to present some short survey on tools and useful results in the ergodic theory of systems which are away from equilibrium. |
Wednesday, 5 November 2025 |
13:00 |
HXLY 130 |
| Jingdong Zhang (Imperial College London) | Artificial neural networks in modulating complex dynamics: Theory and ApplicationAbstract: Dynamical systems are pervasive in nature and engineering, appearing in areas such as neuroscience, life sciences, and molecular dynamics. However, due to nonlinear interactions, network structures, stochastic perturbations, and time delays, these systems often exhibit highly intricate behaviours that are difficult to control or stabilize. This work integrates control and stability theories with neural network architectures to modulate the evolution of complex dynamical systems. Specifically, it develops theoretical frameworks and algorithms for regulating steady states and synchronizing stochastic and delayed networked systems. The study addresses three key challenges in applying neural networks to dynamical systems: the lack of theoretical guarantees, high computational cost, and difficulty in controlling underactuated systems. The proposed methods achieve improved convergence, energy efficiency, and robustness compared to existing state-of-the-art approaches. Finally, the framework is extended to generative modelling, demonstrating its potential application for controlling high-dimensional dynamical processes. |
Wednesday, 12 November 2025 |
14:00 |
HXLY 130 |
| Izaak Neri (King's College London) | Stability of linear systems on sparse random graphs: the role of sign patternsAbstract: Linear stability of dynamical systems is determined by the eigenvalue with the largest real part of Jacobian matrix. For nonsymmetric random graphs, the leading eigenvalue depends strongly on the sign pattern of the weights of the graph. Specifically, for random graphs with sign-antisymmetric weights, the leading eigenvalue is finite in the infinite size limit, while if there is a finite fraction of sign-symmetric weights, then the leading eigenvalue diverges as a function of system size. This result demonstrates the strong stabilising nature of sign-antisymmetric weights. In this seminar, I will first present numerical results on the influence of the sign pattern on the leading eigenvalue, followed by theoretical results explaining these numerical experiments.
[1] A M Mambuca, C Cammarota, and I Neri, Physical Review E (2022)
[2] P Valigi, I Neri, and C Cammarota, Journal of Physics: Complexity (2024)
[3] P Valigi, J Baron, et al., arXiv 2507.20225 (2025). |
Wednesday, 19 November 2025 |
14:00 |
HXLY 130 |
| James Montaldi (University of Manchester) | Equilibria in Contact DynamicsAbstract: TBD |
Wednesday, 26 November 2025 |
14:00 |
HXLY 130 |
DynamIC Workshops and Mini-Courses (Complete List)
| Title |
Date |
Venue |
| One-day workshop on Random Dynamical Systems and Ergodic Theory | Tuesday, 9 September 2025 | HXLY 340, Imperial College London |
| CHAOS (Homoclinic Bifurcations, Strange Attractors, Arnold Diffusion, Fermi Acceleration, Solitons) | Sunday, 24 September 2023 – Friday, 29 September 2023 | Nesin Math Village, Izmir, Turkey |
Short-term DynamIC Visitors (Complete List)
| Name | Affiliation | Arrival | Departure | Host | | Eugenio Pozzoli | Université de Rennes | Monday, 3 November 2025 | Friday, 7 November 2025 | Dmitry Turaev |
Links
|
