Decoding the deterministic nature of black hole IGR J17091-3624

TL;DR

This study investigates the non-linear dynamics of black hole IGR J17091-3624, employing denoising techniques to reveal underlying deterministic behavior previously obscured by noise, aligning it more closely with similar black hole systems.

Contribution

The paper introduces a comprehensive denoising approach using Autoencoder, PCA, SVD, and CI to uncover deterministic signatures in noisy black hole data, challenging prior conclusions of stochasticity.

Findings

Signs of determinism found after denoising

Supports similarity between IGR J17091-3624 and GRS 1915+105

Noise mitigation techniques are effective in revealing underlying dynamics

Abstract

The differentiation between chaotic and stochastic systems has long been scrutinized, particularly in observations where data is often noise-contaminated and finite. Our research examines the dual nature of the black hole X-ray binary IGR J17091-3624, an object whose behavior has been closely studied in parallel to GRS 1915+105. Remarkable similarities in the temporal classes of these two objects are explored in literature. However, this was not the case with their non-linear dynamics: GRS 1915+105 shows signs of determinism, while IGR J17091-3624 was found to be stochastic. In this study, we confront the inherent challenge of noise contamination, as in IGR J17091-3624, faced by previous studies, particularly Poisson noise, which adversely impacts the reliability of non-linear results. We employ several denoising techniques to mitigate noise effects and employ methods like Autoencoder, Principal Component Analysis (PCA), Singular Value Decomposition (SVD), and Correlation Integral (CI) to isolate the deterministic signatures. We have found signs of determinism in IGR J17091-3624 after denoising, thus supporting the hypothesis of it being similar to GRS 1915+105, even as a dynamical system.