En el VIII Congreso de Jóvenes Investigadores de la RSME, se presentarán un total de 21 pósteres. Las descripciones (en inglés) pueden encontrarse en esta página ordenados por orden alfabético de cada presentador. Los pósteres (en inglés) estarán disponibles para su visionado en el Hall del Aula Magna, mientras que la presentación será en la sesión 1 (durante la recepción de bienvenida) o 2 (durante el café de la mañana del jueves), según indica el código de cada póster.
It is known that if $\alpha$ is a Pisot number and $\xi >0$ lies in $\mathbb{Q}(\alpha)$ then the sequence $(\xi \alpha^n)_n$ has finitely many limit points modulo 1. This raises the question of whether it is possible to construct such sequences with a prescribed number of limit points, a problem closely linked to identifying linear recurrences with a prefixed number of residues modulo an integer. Recent results on this problem are presented.
Joint work with Carlo Sanna.
Three balls inequalities are a useful tool in the study of unique continuation properties in the continuum. Our goal is to extend these inequalities to certain discrete lattices, known as periodic graphs. Periodic graphs are graphs in $\mathbb{R}^d$ that remain invariant under translations of particular vectors. We prove that such inequalities holds for Schrödinger operators on a family of periodic graph, and for Laplace operators on a wider family.
Joint work with Aingeru Fernandez Bertolin and Philippe Jaming.
We propose a novel linear probing method for deep transfer learning using minimax risk classifiers (MRCs). This approach ensures robust learning with tight performance bounds. Additionally, we introduce a faster, effective model selection procedure leveraging MRC-provided upper bound, offering a more robust alternative to traditional methods.
Joint work with Santiago Mazuelas.
We propose a new procedure of integration for a class of Liénard I-type second-order equations based on the construction of an associated $C^\infty$-structure using a Lie point symmetry and a $C^\infty$-symmetry of the equation. As an application, we obtain a two-parametric family of travelling wave solutions to a nonlinear partial differential equation which models weakly nonlinear phenomena in a media with convection, diffusion and dissipation.
Joint work with Adrián Ruiz Serván and Mª Concepción Muriel Patino.
Greedy algorithms are a fundamental category of algorithms in mathematics and computer science, characterized by their iterative, locally optimal decision-making approach, which aims to find global optima. In this poster, we will discuss and pose some open questions about two greedy algorithms: the so-called Relaxed Greedy Algorithm and the Thresholding Greedy Algorithm.
A central question in transformation group theory is determining which finite groups act effectively on a given manifold, and a full answer remains elusive. Two main results describe finite abelian groups act on a manifold: the Mann-Su and the CarlssonBaumgartner theorems. This poster introduces the discrete degree of symmetry, an invariant for effective actions of sequences of finite abelian groups of increasing size, analogous to the toral degree of symmetry.
We find a solution to the problem of realizing (finitely generated) permutation modules in topological spaces. This can be seen as a combination of the Kanh's realizability problem for abstract groups and the G-Moore space problem, as it involves realizing a (finite) group G as the group of self homotopy equivalences of a space X, a module M as its integral homology and the action of the group G on such module.
Joint work with Cristina Costoya and Antonio Viruel.
We study the Clifford hierarchy from a group-theoretic perspective, deducing general properties about the cardinality of its levels. These are applicable not just to the usual Clifford hierarchy, but to any analogous construction. With our approach, it is possible to easily deduce results like the finitude of all the levels of general Clifford hierarchies, showing that a bounded number of applications of gate teleporataion schemes can only yield a finite number of fault-tolerant gates.
Joint work with Elías F. Combarro and Ignacio F. Rúa.
In this poster, we extend families of the so-called continuous triangular norms to define families of continuous uninorms within the open square $(0,1)^2$. By leveraging boundary conditions and parameterizations of triangular norms, we construct uninorms that satisfy key aggregation properties. This approach yields new families of uninorms with distinct behaviors, opening avenues for novel aggregation operators.
Work supervised by Marisol Gómez and Raúl Pérez-Fernández.
In this poster we will collect different Brunn-Minkowski type inequalities for a general class of functionals, defined on the family of convex bodies, when dealing with the p-sum of the sets involved.
As a particular case of our approach, we will derive new Lp Brunn-Minkowski type inequalities for both the standard Gaussian measure and the so-called Wills functional, tools of high interest in Convex Geometric Analysis, as shown by the amount of different recent research works on them.
Joint work with Jesús Yepes Nicolás.
In real algebraic geometry, the archimedeanity of quadratic modules is crucial for results like Putinar’s Positivstellensatz. Only theoretical characterizations of this property exist so far. Our approach provides a computational approximation, yielding a concrete decomposition within the quadratic module when it is archimedean. Optimized through parallelization, our implementation has been tested on the MareNostrum 5 supercomputer at the Barcelona Supercomputing Center.
Joint work with Carlos D'Andrea.
Mass-action networks have semialgebraic steady state varieties. We introduce maximal invariant polyhedral supports (MIPS) and we prove that preclusters are dual to maximal invariant polyhedral supports. Given the close relation between MIPS and siphons, we conjecture that siphons and clusters are dual objects, which, we believe, based on the recent work of Craciun et al., might lead to the absence of boundary steady states in toric systems with small codimensional invariant polyhedra.
Joint work with Lamprini Ananiadi.
The Median-of-Means (MoM) estimator is a widely used robust estimator of the mean that is known to achieve the (minimax) optimal estimation error order for heavy-tailed distributions. However, its optimality under adversarial contamination was unknown. In this work, we prove that MoM is optimal in different distributions classes under adversarial contamination. We also identify classes of distributions where MoM becomes suboptimal.
Joint work with Santiago Mazuelas.
The derived category of coherent sheaves encodes important geometric information of a variety. Stability conditions provide the notion of polarization on triangulated categories and have been built in the derived category of projective surfaces and some threefolds. However, constructing and describing the space of these stability conditions remains a challenging problem. In this poster we review the actual state of the art.
We dig into a variety of quantum algorithms for black-box groups, rings and magmas, and their complexity. A key result is proved: given a generating set for a subring, we extend it to an additive generating set in polynomial time. This allows several subring problems to be reduced to already solved abelian group problems. The ideality of a subring is also addressed. For magmas, we analyze existence problems and a generalization of the Deutsch-Jozsa algorithm that remains deterministic.
Joint work with Elías F. Comabrro and Ignacio F. Rúa.
Implicit generative models often struggle with unstable training and mode-dropping when using adversarial discriminators, even in simple 1D cases. This work introduces a novel approach for training 1D implicit models, applying a loss function that measures the discrepancy between the model's samples and a uniform distribution. We extend this method to handle univariate and multivariate temporal data, offering a solution for complex distributions and addressing GAN mode collapse challenges.
Joint work with Pablo Olmos, Manuel Alberto Vázquez and Joaquín Míguez.
This poster reviews the Fast Fourier Transform (FFT) algorithm, focusing on its impact on signal processing and computation. Based on the Cooley-Tukey paper, which introduced an efficient method for calculating the Discrete Fourier Transform (DFT) and its inverse (IDFT), it highlights the FFT's role in reducing computational complexity. The presentation covers the algorithm’s history, compares direct DFT and FFT-1D complexity.
Hyperbolic surfaces have a natural metric inherited from the hyperbolic metric on $\mathbb{H}$, so we can embed metric discs on them and study the maximum radius of the disc $k$-packings embedded on them. In this poster, we introduce an upper bound for the radius of a $k$-packing on a complete, finite-area hyperbolic surface and then, we comment on a couple of characterizations of $k$-extremal surfaces, that is, hyperbolic surfaces that admit a $k$-packing whose radius attains this upper bound.
Joint work with Ernesto Girondo.
Dynamic systems often involve states in complex spaces like Lie groups. This work explores using particle filters to infer states in nonlinear systems, where probability densities are represented by particles. The have challenges, including resampling, solved by an improved method using the Laplace approximation. The generalization of particle filters to Lie groups is also briefly explained.
Consider a polynomial system $f_1(x)=0,\ldots,f_n(x)=0$ of $n$ real polynomials in $n$ variables, where each $f_k$ has a prescribed set $A_k\subseteq \mathbb{Z}^n$ of exponent vectors of size $t_k$. In this poster, we will present a probabilistic version of Kushnirenko's conjecture. We will show that if the coefficients are independent Gaussians of any variance, then the expected number of positive zeros of the random system is bounded from above by $4^{-n} \prod_{k=1}^n t_k(t_k-1)$.
Joint work with Alperen A. Ergür and Maté L. Telek.