arXiv subjects

Yoshitaka Saiki

Publications and source records attributed to Yoshitaka Saiki.

At least 19 recordsLinked to original sources

arXiv2026

Finite Invariant Sets with Bridging Points in Logistic IFS

We investigate iterated function systems (IFS) that randomly alternate between two non-identical one-dimensional maps as simple models of regime-switching dynamical systems. Our primary focus is on finite invariant sets exhibiting ``toss-and-catch'' dynamics, in which trajectories alternate between fixed points and periodic points of the constituent maps. Using two representative types of low-dimensional nonlinear systems--a pair of logistic maps and a combination of logistic and tent maps--we derive exact parameter conditions for several toss-and-catch structures. The comparison between these systems reveals two distinct mechanisms for finite invariant sets: one mediated by bridging points that connect invariant structures of different maps, and another generated by nontrivial intersections shared by the maps themselves. Notably, we identify invariant sets containing bridging points that are not periodic points of either constituent map. These results demonstrate that switching between simple nonlinear maps can generate invariant structures that do not exist in the constituent systems alone, suggesting a more general mechanism for the emergence of invariant structures in random dynamical systems.

math.DS↗

arXiv2026

Geometric structure of ideal data-driven dynamical model using RfR method

The Gaussian radial function-based Regression (RfR) method is a data-driven modeling approach that utilizes physically understandable variables from scalar time series, constructed using delay coordinates and Gaussian radial basis functions. Even when a model successfully describes an approximate trajectory of the original system, data-driven models rarely reconstruct negative Lyapunov exponents of chaotic dynamics. An ''ideal model'' should reconstruct the dynamical structure, including the negative (physically dominant) Lyapunov exponents. Comparing the ideal model and the non-ideal model, we investigate the geometric structure of the attractor of such models using the Lyapunov exponents and the corresponding Lyapunov vectors. Our investigation suggests that the ideal model reconstructs the original system's attractor as a time-delay embedding. By applying the results, we search for a method to construct an ideal model, which persists against the change in hyperparameters.

nlin.CD↗

arXiv2026

On the attractor in a high-dimensional neural network dynamics of reservoir computing: Lyapunov analysis viewpoint

Recent theoretical developments of reservoir computing have clarified a sufficient condition about which reservoir computing can capture the dynamics of a target system, enabling the reconstruction of dynamical invariants. Even when the condition is relaxed, the reservoir computing is found to succeed in reconstructing time series. In this study, we investigate numerically the dynamical structures underlying the embedding structure by comparing the Lyapunov spectrum of a high-dimensional neural network in a reservoir computing model with that of the actual system. We also compute Lyapunov exponents restricted to the tangent space of the inertial manifold in a high-dimensional neural network. Our results provide numerical evidence that reservoir computing can accurately identify the Lyapunov spectrum of the target system, including all negative exponents.

nlin.CD↗

arXiv2025

Long-term prediction of El Niño-Southern Oscillation using reservoir computing with data-driven realtime filter

In recent years, the application of machine learning approaches to time-series forecasting of climate dynamical phenomena has become increasingly active. It is known that applying a band-pass filter to a time-series data is a key to obtaining a high-quality data-driven model. Here, to obtain longer-term predictability of machine learning models, we introduce a new type of band-pass filter. It can be applied to realtime operational prediction workflows since it relies solely on past time series. We combine the filter with reservoir computing, which is a machine-learning technique that employs a data-driven dynamical system. As an application, we predict the multi-year dynamics of the El Niño-Southern Oscillation with the prediction horizon of 24 months using only past time series.

physics.comp-ph↗

arXiv2024

Data-driven ODE modeling of the high-frequency complex dynamics via a low-frequency dynamics model

In our previous paper [N. Tsutsumi, K. Nakai and Y. Saiki, Chaos 32, 091101 (2022)], we proposed a method for constructing a system of differential equations of chaotic behavior from only observable deterministic time series, which we call the radial function-based regression (RfR) method. However, when the targeted variable's behavior is rather complex, the direct application of the RfR method does not function well. In this study, we propose a novel method of modeling such dynamics, including the high-frequency intermittent behavior of a fluid flow, by considering another variable (base variable) showing relatively simple, less intermittent behavior. We construct an autonomous joint model composed of two parts: the first is an autonomous system of a base variable, and the other concerns the targeted variable being affected by a term involving the base variable to demonstrate complex dynamics. The constructed joint model succeeded in not only inferring a short trajectory but also reconstructing chaotic sets and statistical properties obtained from a long trajectory such as the density distributions of the actual dynamics.

nlin.CD↗

arXiv2024

The dynamics of the heterochaos baker maps

The heterochaos baker maps are piecewise affine maps of the unit square or cube introduced in [Nonlinearity 34, 2021, 5744--5761], to provide a hands-on, elementary understanding of complicated phenomena in systems of large degrees of freedom. We review recent progress on a dynamical systems theory of the heterochaos baker maps, and present new results on properties of measures of maximal entropy and the underlying Lebesgue measure. We address several conjectures and questions that may illuminate new aspects of heterochaos and inspire future research.

math.DS↗

arXiv2024

A laminar chaotic saddle within a turbulent attractor

Intermittent switchings between weakly chaotic (laminar) and strongly chaotic (bursty) states are often observed in systems with high-dimensional chaotic attractors, such as fluid turbulence. They differ from the intermittency of a low-dimensional system accompanied by the stability change of a fixed point or a periodic orbit in that the intermittency of a high-dimensional system tends to appear in a wide range of parameters. This paper considers a case where the skeleton of a laminar state $L$ exists as a proper chaotic subset $S$ of a chaotic attractor $X$, that is, $S\ \subsetneq\ X$. We characterize such a laminar state $L$ by a chaotic saddle $S$, which is densely filled with periodic orbits of different numbers of unstable directions. This study demonstrates the presence of chaotic saddles underlying intermittency in fluid turbulence and phase synchronization. Furthermore, we confirm that chaotic saddles persist for a wide range of parameters. Also, a kind of phase synchronization turns out to occur in the turbulent model.

nlin.CD↗

arXiv2024

Data-driven modeling from biased small training data using periodic orbits

In this study, we investigate the effect of reservoir computing training data on the reconstruction of chaotic dynamics. Our findings indicate that a training time series comprising a few periodic orbits of low periods can successfully reconstruct the Lorenz attractor. We also demonstrate that biased training data does not negatively impact reconstruction success. Our method's ability to reconstruct a physical measure is much better than the so-called cycle expansion approach, which relies on weighted averaging. Additionally, we demonstrate that fixed point attractors and chaotic transients can be accurately reconstructed by a model trained from a few periodic orbits, even when using different parameters.

physics.data-an↗

arXiv2023

Constructing low-dimensional ordinary differential equations from chaotic time series of high/infinite-dimensional systems using radial function-based regression

In our previous study (N. Tsutsumi, K. Nakai and Y. Saiki (2022)) we proposed a method of constructing a system of differential equations of chaotic behavior only from observable deterministic time series, which we will call radial function-based regression (RfR) method. The RfR method employs a regression using Gaussian radial basis functions together with polynomial terms to facilitate the robust modeling of chaotic behavior. In this paper, we apply the RfR method to several types of relatively high-dimensional deterministic time series generated by a partial differential equation, a delay differential equation, a turbulence model, and intermittent dynamics. The case when the observation includes noise is also tested. We have effectively constructed a system of differential equations for each of these examples, which is assessed from the point of view of time series forecast, reconstruction of invariant sets, and invariant densities. We find that in some of the models, an appropriate trajectory is realized on the chaotic saddle and is identified by the Stagger-and-Step method.

math.DS↗

arXiv2023

Characterizing Data Assimilation in Navier-Stokes Turbulence with Transverse Lyapunov Exponents

Data assimilation (DA) reconstructing small-scale turbulent structures is crucial for forecasting and understanding turbulence. This study proposes a theoretical framework for DA based on ideas from chaos synchronization, in particular, the transverse Lyapunov exponents (TLEs). The analysis with TLEs characterizes a critical length scale, below which the turbulent dynamics is synchronized to the larger-scale turbulent dynamics, indicating successful DA. An underlying link between TLEs and the maximal Lyapunov exponent suggests that the critical length scale depends on the Reynolds number. Furthermore, we discuss new directions of DA algorithms based on the proposed framework.

physics.flu-dyn↗

arXiv2023

Hausdorff dimension of Cantor intersections and robust heterodimensional cycles for heterochaos horseshoe maps

As a model to provide a hands-on, elementary understanding of chaotic dynamics in dimension three, we introduce a $C^2$-open set of diffeomorphisms of $\mathbb R^3$ having two horseshoes with different dimensions of instability. We prove that: the unstable set of one horseshoe and the stable set of the other are of Hausdorff dimension nearly $2$ whose cross sections are Cantor sets; the intersection of the unstable and stable sets contains a fractal set of Hausdorff dimension nearly $1$. As a corollary we detect $C^2$-robust heterodimensional cycles. Our proof employs the theory of normally hyperbolic invariant manifolds and the thicknesses of Cantor sets.

math.DS↗

arXiv2023

The twisted baker map

As a model to provide a hands-on, elementary understanding of "vortex dynamics", we introduce a piecewise linear non-invertible map called a twisted baker map. We show that the set of hyperbolic repelling periodic points with complex conjugate eigenvalues and that without complex conjugate eigenvalues are simultaneously dense in the phase space. We also show that these two sets equidistribute with respect to the normalized Lebesgue measure, in spite of a non-uniformity in their Lyapunov exponents.

math.DS↗

arXiv2022

Machine learning prediction of the MJO extends beyond one month

The prediction of the Madden-Julian Oscillation (MJO), a massive tropical weather event with vast global socio-economic impacts, has been infamously difficult with physics-based weather prediction models. Here we construct a machine learning model using reservoir computing technique that forecasts the real-time multivariate MJO index (RMM), a macroscopic variable that represents the state of the MJO. The training data was refined by developing a novel filter that extracts the recurrency of MJO signals from the raw atmospheric data and selecting a suitable time-delay coordinate of the RMM. The model demonstrated the skill to forecast the state of MJO events for a month from the pre-developmental stages. Best-performing cases predicted the RMM sequence over two months, which exceeds the expected inherent predictability limit of the MJO.

physics.ao-ph↗

arXiv2022

Piecewise-linear maps with heterogeneous chaos

Chaotic dynamics can be quite heterogeneous in the sense that in some regions the dynamics are unstable in more directions than in other regions. When trajectories wander between these regions, the dynamics is complicated. We say a chaotic invariant set is heterogeneous when arbitrarily close to each point of the set there are different periodic points with different numbers of unstable dimensions. We call such dynamics heterogeneous chaos (or hetero-chaos), While we believe it is common for physical systems to be hetero-chaotic, few explicit examples have been proved to be hetero-chaotic. Here we present two more explicit dynamical systems that are particularly simple and tractable with computer. It will give more intuition as to how complex even simple systems can be. Our maps have one dense set of periodic points whose orbits are 1D unstable and another dense set of periodic points whose orbits are 2D unstable. Moreover, they are ergodic relative to the Lebesgue measure.

math.DS↗

arXiv2022

Synchronization analysis between exchange rates on the basis of purchasing power parity using the Hilbert transform

Synchronization is a phenomenon in which a pair of fluctuations adjust their rhythms when interacting with each other. We measure the degree of synchronization between the U.S. dollar (USD) and euro exchange rates and between the USD and Japanese yen exchange rates on the basis of purchasing power parity (PPP) over time. We employ a method of synchronization analysis using the Hilbert transform, which is common in the field of nonlinear science. We find that the degree of synchronization is high most of the time, suggesting the establishment of PPP. The degree of synchronization does not remain high across periods with economic events with asymmetric effects, such as the U.S. real estate bubble.

econ.EM↗

arXiv2022

Regional Synchronization during Economic Contraction: The Case of the U.S. and Japan

Two decades of studies have found significant regional differences in the timing of transitions in national business cycles and their durations. Earlier studies partly detect regional synchronization during business cycle expansions and contractions in Europe, the United States, and Japan. We examine this possibility applying a sophisticated method for identifying the time-varying degree of synchronization to regional business cycle data in the U.S. and Japan. The method is prominent in nonlinear sciences but has been infrequently applied in business cycle studies.We find that synchronization in regional business cycles increased during contractions and decreased during expansions throughout the period under study.Such asymmetry between the contraction and expansion phases of a business cycle will contribute our better understanding of the phenomenon of business cycles.

econ.GN↗

arXiv2022

Constructing differential equations using only a scalar time-series about continuous time chaotic dynamics

We propose a simple method of constructing a system of differential equations of chaotic behavior based on the regression only from a scalar observable time-series data. The estimated system enables us to reconstruct invariant sets and statistical properties as well as to infer short time-series. Our successful modeling relies on the introduction of a set of Gaussian radial basis functions to capture local structure. The proposed method is used to construct a system of ordinary differential equations whose orbit reconstructs a time-series of a variable of the well-known Lorenz system as a simple but typical example. A system for a macroscopic fluid variable is also constructed.

math.DS↗

arXiv2021

Dynamical system analysis of a data-driven model constructed by reservoir computing

This study evaluates data-driven models from a dynamical system perspective, such as unstable fixed points, periodic orbits, chaotic saddle, Lyapunov exponents, manifold structures, and statistical values. We find that these dynamical characteristics can be reconstructed much more precisely by a data-driven model than by computing directly from training data. With this idea, we predict the laminar lasting time distribution of a particular macroscopic variable of chaotic fluid flow, which cannot be calculated from a direct numerical simulation of the Navier-Stokes equation because of its high computational cost.

math.DS↗