Juan-Pablo Ortega Lahuerta
Last Name
Ortega Lahuerta
First name
Juan-Pablo
Email
juan-pablo.ortega@unisg.ch
Phone
+41 71 224 2476
Web Site
30 results
Now showing 1 - 10 of 30
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Reservoir Kernels and Volterra Series(2025); ; Type:journal articleJournal:IEEE Transactions on Neural Networks and Learning SystemsScopus© Citations 2 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Memory of recurrent networks: Do we compute it right?Numerical evaluations of the memory capacity (MC) of recurrent neural networks reported in the literature often contradict well-established theoretical bounds. In this paper, we study the case of linear echo state networks, for which the total memory capacity has been proven to be equal to the rank of the corresponding Kalman controllability matrix. We shed light on various reasons for the inaccurate numerical estimations of the memory, and we show that these issues, often overlooked in the recent literature, are of an exclusively numerical nature. More explicitly, we prove that when the Krylov structure of the linear MC is ignored, a gap between the theoretical MC and its empirical counterpart is introduced. As a solution, we develop robust numerical approaches by exploiting a result of MC neutrality with respect to the input mask matrix. Simulations show that the memory curves that are recovered using the proposed methods fully agree with the theory.Type:journal articleJournal:Journal of Machine Learning Research - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Reservoir computing for macroeconomic forecasting with mixed-frequency data(2024-07); ; Scopus© Citations 20 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Scopus© Citations 12 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Memory of recurrent networks: Do we compute it right?(2023-05-02); ; Numerical evaluations of the memory capacity (MC) of recurrent neural networks reported in the literature often contradict well-established theoretical bounds. In this paper, we study the case of linear echo state networks, for which the total memory capacity has been proven to be equal to the rank of the corresponding Kalman controllability matrix. We shed light on various reasons for the inaccurate numerical estimations of the memory, and we show that these issues, often overlooked in the recent literature, are of an exclusively numerical nature. More explicitly, we prove that when the Krylov structure of the linear MC is ignored, a gap between the theoretical MC and its empirical counterpart is introduced. As a solution, we develop robust numerical approaches by exploiting a result of MC neutrality with respect to the input mask matrix. Simulations show that the memory curves that are recovered using the proposed methods fully agree with the theory.Type:journal article - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Approximation bounds for random neural networks and reservoir systems(Institute of Mathematical Statistics, 2023-02); ; Type:journal articleJournal:The Annals of Applied ProbabilityVolume:33Issue:1 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Tracing curves in the plane: geometric-invariant learning from human demonstrations(2023); ; Campolo, DomenicoType:journal article - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Discrete-time signatures and randomness in reservoir computing(2021); ; ; Teichmann, JosefA new explanation of geometric nature of the reservoir computing phenomenon is presented. Reservoir computing is understood in the literature as the possibility of approximating input/output systems with randomly chosen recurrent neural systems and a trained linear readout layer. Light is shed on this phenomenon by constructing what is called strongly universal reservoir systems as random projections of a family of state-space systems that generate Volterra series expansions. This procedure yields a state-affine reservoir system with randomly generated coefficients in a dimension that is logarithmically reduced with respect to the original system. This reservoir system is able to approximate any element in the fading memory filters class just by training a different linear readout for each different filter. Explicit expressions for the probability distributions needed in the generation of the projected reservoir system are stated and bounds for the committed approximation error are provided.Type:journal articleScopus© Citations 33 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Fading memory echo state networks are universal(2021-06); Type:journal articleJournal:Neural NetworksScopus© Citations 73 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Dimension reduction in recurrent networks by canonicalizationMany recurrent neural network machine learning paradigms can be formulated using state-space representations. The classical notion of canonical state-space realization is adapted in this paper to accommodate semi-infinite inputs so that it can be used as a dimension reduction tool in the recurrent networks setup. The so-called input forgetting property is identified as the key hypothesis that guarantees the existence and uniqueness (up to system isomorphisms) of canonical realizations for causal and time-invariant input/output systems with semi-infinite inputs. Additionally, the notion of optimal reduction coming from the theory of symmetric Hamiltonian systems is implemented in our setup to construct canonical realizations out of input forgetting but not necessarily canonical ones. These two procedures are studied in detail in the framework of linear fading memory input/output systems. {Finally, the notion of implicit reduction using reproducing kernel Hilbert spaces (RKHS) is introduced which allows, for systems with linear readouts, to achieve dimension reduction without the need to actually compute the reduced spaces introduced in the first part of the paper.Type:journal articleJournal:Journal of Geometric MechanicsVolume:13Issue:4Scopus© Citations 18
- «
- 1 (current)
- 2
- 3
- »