arXiv subjects

Ana Mancic

Publications and source records attributed to Ana Mancic.

4 recordsLinked to original sources

arXiv2026

Lyapunov Spectral Analysis of Speech Embedding Trajectories in Psychosis

We analyze speech embeddings from structured clinical interviews of psychotic patients and healthy controls by treating language production as a high-dimensional dynamical process. Lyapunov exponent (LE) spectra are computed from word-level and answer-level embeddings generated by two distinct large language models, allowing us to assess the stability of the conclusions with respect to different embedding presentations. Word-level embeddings exhibit uniformly contracting dynamics with no positive LE, while answer-level embeddings, in spite of the overall contraction, display a number of positive LEs and higher-dimensional attractors. The resulting LE spectra robustly separate psychotic from healthy speech, while differentiation within the psychotic group is not statistically significant overall, despite a tendency of the most severe cases to occupy distinct dynamical regimes. These findings indicate that nonlinear dynamical invariants of speech embeddings provide a physics-inspired probe of disordered cognition whose conclusions remain stable across embedding models.

nlin.AO↗

arXiv2026

Thermalization in a nonlinear variant of the discrete nonlinear Schr\"odinger Equation

We study the thermalization properties of a fully nonlinear lattice model originally derived from the two-dimensional cubic defocusing nonlinear Schr{\"o}dinger equation (NLS) using analytical and numerical methods. The model conserves both energy and norm, whose densities define a microcanonical energy-norm parameter space, while the nonlinear nearest-neighbor coupling is controlled by a parameter $D$. Within this space, our analysis identifies broad parameter regimes in which the dynamics is ergodic not only within but also outside the standard Gibbs region, indicating the need for a modified statistical description. At higher energies, the system instead exhibits long-lived compacton-mediated localization and signatures of weak nonergodicity, as evidenced by finite-time variances, excursion-time statistics, and probability distributions of local amplitudes. We show that stronger coupling $D$ enhances fluctuations and accelerates the crossover of the finite-time variance of the local norm density from an initial $\sim T^{-1/2}$ decay toward the faster $\sim T^{-1}$ decay characteristic of ergodic thermalization, where $T$ denotes the averaging time. In the high-energy regime, weak coupling $D\leq 1$ favors persistent single-site compacton localization, whereas stronger coupling $D>1$ yields long-lived two-site localization. Our results provide insights into the interplay between thermalization, localization, and non-Gibbs statistical behavior in genuinely nonlinear systems.

cond-mat.stat-mech↗

arXiv2020

Deep learning-based classification of high intensity patterns in photorefractive crystals

In this paper, we establish a new scheme for identification and classification of high intensity events generated by the propagation of light through a photorefractive SBN crystal. Among these events, which are the inevitable consequence of the development of modulation instability, are speckling and soliton-like patterns. The usual classifiers developed on statistical measures, such as the significant intensity, often provide only a partial characterization of these events. Here, we try to overcome this deficiency by implementing the convolution neural network method to relate experimental data of light intensity distribution and corresponding numerical outputs with different high intensity regimes. The train and test sets are formed of experimentally obtained intensity profiles at the crystal output facet and corresponding numerical profiles. The accuracy of detection of speckles reaches maximum value of 100%, while the accuracy of solitons and caustic detection is above 97%. These performances are promising for the creation of neural network based routines for prediction of extreme events in wave media.

nlin.PS↗

arXiv2004

Moving weakly relativistic electromagnetic solitons in laser-plasmas

A case of moving one-dimensional electromagnetic (EM) solitons formed in a relativistic interaction of a linearly polarized laser light with underdense cold plasma is investigated. The relativistic Lorentz force in an intense laser light pushes electrons into longitudinal motion generating coupled longitudal-transverse wave modes. In a weakly relativistic approximation these modes are well described by a dynamical equation of the generalized nonlinear Schrodinger type, with two additional nonlocal terms [1]. An original analytical solution for a moving EM soliton case is here calculated in an implicit form. The soliton motion down-shifts the soliton eigen-frequency while decreases its amplitude. An influence of the soliton velocity on stability properties is analytically predicted.

physics.plasm-ph↗