-
Reverse-BSDE Monte Carlo
Authors:
Jairon H. N. Batista,
Flávio B. Gonçalves,
Yuri F. Saporito,
Rodrigo S. Targino
Abstract:
Recently, there has been a growing interest in generative models based on diffusions driven by the empirical robustness of these methods in generating high-dimensional photorealistic images and the possibility of using the vast existing toolbox of stochastic differential equations. %This remarkable ability may stem from their capacity to model and generate multimodal distributions. In this work, w…
▽ More
Recently, there has been a growing interest in generative models based on diffusions driven by the empirical robustness of these methods in generating high-dimensional photorealistic images and the possibility of using the vast existing toolbox of stochastic differential equations. %This remarkable ability may stem from their capacity to model and generate multimodal distributions. In this work, we offer a novel perspective on the approach introduced in Song et al. (2021), shifting the focus from a "learning" problem to a "sampling" problem. To achieve this, we reformulate the equations governing diffusion-based generative models as a Forward-Backward Stochastic Differential Equation (FBSDE), which avoids the well-known issue of pre-estimating the gradient of the log target density. The solution of this FBSDE is proved to be unique using non-standard techniques. Additionally, we propose a numerical solution to this problem, leveraging on Deep Learning techniques. This reformulation opens new pathways for sampling multidimensional distributions with densities known up to a normalization constant, a problem frequently encountered in Bayesian statistics.
△ Less
Submitted 10 May, 2025;
originally announced May 2025.
-
Synchrony and canards in two coupled FitzHugh--Nagumo equations
Authors:
Bruno F. F. Gonçalves,
Isabel S. Labouriau,
Alexandre A. P. Rodrigues
Abstract:
We describe the fast-slow dynamics of two FitzHugh--Nagumo equations coupled symmetrically through the slow equations.
We use symmetry arguments to find a non-empty open set of parameter values for which the two equations synchronise, and another set with antisynchrony -- where the solution of one equation is minus the solution of the other.
By combining the dynamics within the synchrony and a…
▽ More
We describe the fast-slow dynamics of two FitzHugh--Nagumo equations coupled symmetrically through the slow equations.
We use symmetry arguments to find a non-empty open set of parameter values for which the two equations synchronise, and another set with antisynchrony -- where the solution of one equation is minus the solution of the other.
By combining the dynamics within the synchrony and antisynchrony subspaces, we also obtain bistability -- where these two types of solution coexist as hyperbolic attractors. They persist under small perturbation of the parameters.
Canards are shown to give rise to mixed-mode oscillations.
They also initiate small amplitude transient oscillations before the onset of large amplitude relaxation oscillations. We also discuss briefly the effect of asymmetric coupling, with periodic forcing of one of the equations by the other. We illustrate our results with numerical simulations.
△ Less
Submitted 23 July, 2025; v1 submitted 16 March, 2025;
originally announced March 2025.
-
Bifurcations and canards in the FitzHugh-Nagumo system: a tutorial in fast-slow dynamics
Authors:
Bruno F. F. Gonçalves,
Isabel S. Labouriau,
Alexandre A. P. Rodrigues
Abstract:
In this article, we study the FitzHugh-Nagumo $(1,1)$--fast-slow system where the vector fields associated to the slow/fast equations come from the reduction of the Hodgin-Huxley model for the nerve impulse. After deriving dynamical properties of the singular and regular cases, we perform a bifurcation analysis and we investigate how the parameters (of the affine slow equation) impact the dynamics…
▽ More
In this article, we study the FitzHugh-Nagumo $(1,1)$--fast-slow system where the vector fields associated to the slow/fast equations come from the reduction of the Hodgin-Huxley model for the nerve impulse. After deriving dynamical properties of the singular and regular cases, we perform a bifurcation analysis and we investigate how the parameters (of the affine slow equation) impact the dynamics of the system. The study of codimension one bifurcations and the numerical locus of canards concludes this case-study. All theoretical results are numerically illustrated.
△ Less
Submitted 1 May, 2025; v1 submitted 17 November, 2024;
originally announced November 2024.
-
On the dynamics of rotating rank-one strange attractors families
Authors:
Alexandre A. P. Rodrigues,
Bruno F. Gonçalves
Abstract:
In this article, we study a two-parameter family of rotating rank-one maps defined on $\textbf{S}^1\times [1, 1+b]\times \textbf{S}^1$, with $b\gtrsim 0$, whose dynamics is characterised by a coupling of a family of planar maps exhibiting rank-one strange attractors and an Arnold family of circle maps. The main result is about the dynamics on the skew-product, which is governed by the existence an…
▽ More
In this article, we study a two-parameter family of rotating rank-one maps defined on $\textbf{S}^1\times [1, 1+b]\times \textbf{S}^1$, with $b\gtrsim 0$, whose dynamics is characterised by a coupling of a family of planar maps exhibiting rank-one strange attractors and an Arnold family of circle maps. The main result is about the dynamics on the skew-product, which is governed by the existence and prevalence of strange attractors in the corresponding resonance tongues of the Arnold family. The strange attractors carry the unique physical measure of the system, which determines the behaviour of Lebesgue-almost all initial conditions. This phenomenon can be considered as the transition dynamics from a strange attractor with one positive Lyapunov exponent to hyperchaos. Besides an analytical rigorous proof, we illustrate the main results with numerical simulations. We also conjecture how persistent hyperchaos can be obtained.
△ Less
Submitted 19 August, 2024;
originally announced August 2024.
-
Perfect simulation for interacting Hawkes processes with variable length memory
Authors:
Branda Goncalves,
Paul Gresland
Abstract:
We consider a nonlinear multivariate Hawkes process having a variable length memory which allows to describe the activity of a neuronal network by its membrane potential. We propose a graphical construction of the process and we construct, by means of a perfect simulation algorithm, a stationary version of the process. By making the hypothesis that the spiking rate $β_i$ of the neuron $i \in I $ i…
▽ More
We consider a nonlinear multivariate Hawkes process having a variable length memory which allows to describe the activity of a neuronal network by its membrane potential. We propose a graphical construction of the process and we construct, by means of a perfect simulation algorithm, a stationary version of the process. By making the hypothesis that the spiking rate $β_i$ of the neuron $i \in I $ is bounded, we construct an algorithm based on a priori realizations of the Poisson process $(M^i, i \in I)$. We show that there exists a critical value $δ_c$ such that if $\underlineδ > δ_c$ (where $\underlineδ= \inf_i{δ_i}$ with $δ_i = \frac{β_{i* }}{β^*_i-β_{i*}}$ ) the process is ergodic.
△ Less
Submitted 19 September, 2022;
originally announced September 2022.
-
An Interacting Neuronal Network with Inhibition: theoretical analysis and perfect simulation
Authors:
Branda Goncalves
Abstract:
We study a purely inhibitory neural network model where neurons are represented by their state of inhibition. The study we present here is partially based on the work of Cottrell \cite{Cot} and Fricker et al. \cite{FRST}. The spiking rate of a neuron depends only on its state of inhibition. When a neuron spikes, its state is replaced by a random new state, independently of anything else and the in…
▽ More
We study a purely inhibitory neural network model where neurons are represented by their state of inhibition. The study we present here is partially based on the work of Cottrell \cite{Cot} and Fricker et al. \cite{FRST}. The spiking rate of a neuron depends only on its state of inhibition. When a neuron spikes, its state is replaced by a random new state, independently of anything else and the inhibition state of the other neurons increase by a positive value. Using the Perron-Frobenius theorem, we show the existence of a Lyapunov function for the process. Furthermore, we prove a local Doeblin condition which implies the existence of an invariant measure for the process. Finally, we extend our model to the case where the neurons are indexed by $ \mathbb{Z}. $ We construct a perfect simulation algorithm to show the recurrence of the process under certain conditions. To do this, we rely on the classical contour technique used in the study of contact processes, and assuming that the spiking rate lies on the interval $[ β_* , β^* ], $ we show that there is a critical threshold for the ratio $ δ= \frac{β_*}{β^* - β_*}$ over which the process is ergodic. \\ \textbf{Keywords}: spiking rate, interacting neurons, perfect simulation algorithm, classical contour technique.
△ Less
Submitted 18 October, 2021; v1 submitted 13 October, 2021;
originally announced October 2021.
-
Scaling features of two special Markov chains involving total disasters
Authors:
Branda Goncalves,
Thierry Huillet
Abstract:
Catastrophe Markov chain population models have received a lot of attention in the recent past. We herewith consider two special cases of such models involving total disasters, both in discrete and in continuous-time. Depending on the parameters range, the two models can show up a recurrence/transience transition and, in the critical case, a positive/null recurrence transition. The collapse transi…
▽ More
Catastrophe Markov chain population models have received a lot of attention in the recent past. We herewith consider two special cases of such models involving total disasters, both in discrete and in continuous-time. Depending on the parameters range, the two models can show up a recurrence/transience transition and, in the critical case, a positive/null recurrence transition. The collapse transition probabilities are chosen in such a way that the models are exactly solvable and, in case of positive recurrence, intimately related to the extended Sibuya and Pareto-Zipf distributions whose divisibility and self-decomposability properties are shown relevant. The study includes: existence and shape of the invariant measure, time-reversal, return time to the origin, contact probability at the origin, extinction probability, height and length of the excursions, a renewal approach to the fraction of time spent in the catastrophic state, scale function, first time to collapse and first-passage times, divisibility properties.
△ Less
Submitted 11 January, 2021;
originally announced January 2021.
-
On decay-surge population models
Authors:
Branda Goncalves,
Thierry Huillet,
Eva Löcherbach
Abstract:
We consider continuous space-time decay-surge population models which are semi- stochastic processes for which deterministically declining populations, bound to fade away, are rein- vigorated at random times by bursts or surges of random sizes. In a particular separable framework (in a sense made precise below) we provide explicit formulae for the scale (or harmonic) function and the speed measure…
▽ More
We consider continuous space-time decay-surge population models which are semi- stochastic processes for which deterministically declining populations, bound to fade away, are rein- vigorated at random times by bursts or surges of random sizes. In a particular separable framework (in a sense made precise below) we provide explicit formulae for the scale (or harmonic) function and the speed measure of the process. The behavior of the scale function at infinity allows to formulate conditions under which such processes either explode or are transient at infinity, or Harris recurrent. A description of the structures of both the discrete-time embedded chain and extreme record chain of such continuous-time processes is supplied.
△ Less
Submitted 8 November, 2021; v1 submitted 1 December, 2020;
originally announced December 2020.
-
On population growth with catastrophes
Authors:
Branda Goncalves,
Thierry Huillet,
Eva Löcherbach
Abstract:
In this paper we study a particular class of Piecewise deterministic Markov processes (PDMP's) which are semi-stochastic catastrophe versions of deterministic population growth models. In between successive jumps the process follows a flow describing deterministic population growth. Moreover, at random jump times, governed by state-dependent rates, the size of the population shrinks by a random am…
▽ More
In this paper we study a particular class of Piecewise deterministic Markov processes (PDMP's) which are semi-stochastic catastrophe versions of deterministic population growth models. In between successive jumps the process follows a flow describing deterministic population growth. Moreover, at random jump times, governed by state-dependent rates, the size of the population shrinks by a random amount of its current size, an event possibly leading to instantaneous local (or total) extinction. A special separable shrinkage transition kernel is investigated in more detail, including the case of total disasters. We discuss conditions under which such processes are recurrent (positive or null) or transient. To do so, we introduce a modified scale function which is used to compute, when relevant, the law of the height of excursions and to decide if the process is recurrent or not. The question of the finiteness of the time to extinction is investigated together with the evaluation of the mean time to extinction when the last one is finite. Some information on the embedded jump chain of the PDMP is also required when dealing with the classification of states 0 and infinity that we exhibit.
△ Less
Submitted 8 June, 2021; v1 submitted 7 July, 2020;
originally announced July 2020.
-
On the definition of likelihood function
Authors:
Flávio B. Gonçalves,
Pedro Franklin
Abstract:
We discuss a general definition of likelihood function in terms of Radon-Nikodým derivatives. The definition is validated by the Likelihood Principle once we establish a result regarding the proportionality of likelihood functions under different dominating measures. This general framework is particularly useful when there exists no or more than one obvious choice for a dominating measure as in so…
▽ More
We discuss a general definition of likelihood function in terms of Radon-Nikodým derivatives. The definition is validated by the Likelihood Principle once we establish a result regarding the proportionality of likelihood functions under different dominating measures. This general framework is particularly useful when there exists no or more than one obvious choice for a dominating measure as in some infinite-dimensional models. We discuss the importance of considering continuous versions of densities and how these are related to the Likelihood Principle and the basic concept of likelihood. We also discuss the use of the predictive measure as a dominating measure in the Bayesian approach. Finally, some examples illustrate the general definition of likelihood function and the importance of choosing particular dominating measures in some cases.
△ Less
Submitted 21 June, 2021; v1 submitted 25 June, 2019;
originally announced June 2019.
-
Grassmannians and conformal structure on absolutes
Authors:
Sasha Anan'in,
Eduardo C. Bento Goncalves,
Carlos H. Grossi
Abstract:
We study grassmannians associated with a linear space with a nondegenerate hermitian form. The geometry of these grassmannians allows us to explain the relation between a (pseudo-)riemannian projective geometry and the conformal structure on its ideal boundary (absolute). Such relation encompasses, for instance, the usual conformal structure on the absolute of real hyperbolic space, the usual conf…
▽ More
We study grassmannians associated with a linear space with a nondegenerate hermitian form. The geometry of these grassmannians allows us to explain the relation between a (pseudo-)riemannian projective geometry and the conformal structure on its ideal boundary (absolute). Such relation encompasses, for instance, the usual conformal structure on the absolute of real hyperbolic space, the usual conformal structure on the absolute of de Sitter space, the conformal contact structure on the absolute of complex hyperbolic space, and the causal structure on the absolute of anti-de Sitter space.
△ Less
Submitted 24 January, 2020; v1 submitted 26 July, 2009;
originally announced July 2009.
-
A Hyperelliptic View on Teichmuller Space. I
Authors:
Sasha Anan'in,
Eduardo C. Bento Goncalves
Abstract:
We explicitly describe the Teichmuller space TH_n of hyperelliptic surfaces in terms of natural and effective coordinates as the space of certain (2n-6)-tuples of distinct points on the ideal boundary of the Poincare disc. We essentially use the concept of a simple earthquake which is a particular case of a Fenchel-Nielsen twist deformation. Such earthquakes generate a group that acts transitive…
▽ More
We explicitly describe the Teichmuller space TH_n of hyperelliptic surfaces in terms of natural and effective coordinates as the space of certain (2n-6)-tuples of distinct points on the ideal boundary of the Poincare disc. We essentially use the concept of a simple earthquake which is a particular case of a Fenchel-Nielsen twist deformation. Such earthquakes generate a group that acts transitively on TH_n. This fact can be interpreted as a continuous analog of the well-known Dehn theorem saying that the mapping class group is generated by Dehn twists. We find a simple and effective criterion that verifies if a given representation of the surface group π_1Σin the group of isometries of the hyperbolic plane is faithful and discrete. The article also contains simple and elementary proofs of several known results, for instance, of W. M. Goldman's theorem [Gol1] characterizing the faithful discrete representations as having maximal Toledo invariant (which is essentially the area of the representation in the two-dimensional case).
△ Less
Submitted 9 July, 2009; v1 submitted 11 September, 2007;
originally announced September 2007.
