Aims & Scope

Physica D (Nonlinear Phenomena) publishes research and review articles reporting on theoretical and experimental work, techniques, and ideas that advance the understanding of nonlinear phenomena.

The scope of the journal encompasses mathematical methods for nonlinear systems including: wave motion, pattern formation and collective phenomena in physical, chemical and biological systems; hydrodynamics and turbulence; integrable and Hamiltonian systems; and data-driven dynamical systems. The journal encourages submissions in established and emerging application domains, for example applications of nonlinear science to artificial intelligence, robotics, control theory, complex networks, and social and economic dynamics.

Special Issue on New Horizons at the Intersections of Machine Learning, Dynamical Systems, and Algorithmic Information Theory
Submission Date: 2026-02-28

The evolving landscape of theoretical and applied sciences, particularly in the analysis of complex systems, has brought forth the convergence of three pivotal areas: Machine Learning (ML), Dynamical Systems Theory (DST), and Algorithmic Information Theory (AIT). This special issue aims to explore the rich interfaces between ML, DST, and AIT, fostering a multidisciplinary dialogue that could unveil new methodologies, theoretical insights, and practical applications. By delving into the synergies and applications among these domains, we seek to push the boundaries of current knowledge and provide innovative tools for the analysis and understanding of complex systems.

Background:

Dynamical Systems Theory, rooted in the work of pioneers like Poincaré and Lyapunov, provides a robust mathematical framework for analyzing systems dynamics through models expressed in differential or difference equations. Its application spans complex and chaotic systems, including climate dynamics and biological networks. Conversely, Machine Learning, with its algorithm-driven approaches, excels in environments rich in data but lacking explicit models, offering groundbreaking applications from computer vision to stock market analysis. Algorithmic Information Theory, with its focus on the inherent complexity of data and systems, bridges foundational concepts in computation and information to provide a deeper understanding of both ML models and dynamical systems.

Themes and Objectives:

ML for DST: This theme emphasizes the use of ML techniques to analyze and model dynamical systems based on observed data. It seeks contributions that explore data-driven methodologies to extend classical DST theories, aiming to capture the underlying dynamics of complex systems with high-dimensional data.

DST for ML: Here, the focus is on applying DST principles to enhance the understanding and performance of ML algorithms. Submissions should explore how dynamical systems theory can offer new perspectives on the stability, convergence, and behavior of algorithms under various conditions.

ML for AIT: Contributions under this theme will explore how machine learning methodologies can be employed to tackle problems within Algorithmic Information Theory, such as optimizing for better compression and prediction algorithms that approach Kolmogorov Complexity.

AIT for ML: This area seeks to understand and improve ML algorithms through the lens of AIT, examining the theoretical underpinnings that predict algorithmic performance and limitations. Papers should aim to provide new insights into the applicability and efficiency of ML methods for complex problem-solving.

DST for AIT: To explore how DST can inform and enrich AIT, particularly in areas of complexity and randomness. This theme invites studies that apply dynamical systems' insights to algorithmic questions, offering new perspectives on information content, process complexity, and the generation of complex patterns from simple rules.

AIT for DST: To apply AIT concepts to deepen the understanding of dynamical systems, particularly regarding predictability, information dynamics, and system complexity. Contributions should aim to leverage algorithmic principles to shed light on the behavior of dynamical systems, their information content, and their evolution over time.

Guest editors:

Executive Guest Editor
Dr. Boumediene Hamzi
Organization*: Caltech, Pasadena, CA, USA & The Alan Turing Institute, London, UK
Institutional Email Address*: bhamzi@turing.ac.uk
Areas of Expertise*: Machine Learning, Dynamical Systems, Algorithmic Information Theory

Co-Guest Editors
Dr. Kamal Dingle
Organization*: Gulf Institute of Science and Technology, Kuwait
Institutional Email Address*: Dingle.K@gust.edu.kw
Areas of Expertise*: Algorithmic Information Theory, Theoretical Biology

Prof. Marcus Hutter
Organization*: Google DeepMind, London, UK, and Australian National University, Canberra, Australia
Institutional Email Address*: marcus.hutter@anu.edu.au
Areas of Expertise*: Algorithmic Information Theory, Machine Learning

Dr. Qinxiao Li
Organization*: Department of Mathematics, National University of Singapore, 117543, Singapore.
Institutional Email Address*: qianxiao@nus.edu.sg
Areas of Expertise*: Machine Learning, Dynamical Systems

Dr. Tanya Schmah
Organization*: University of Ottawa, Canada
Institutional Email Address*: tschmah@uottawa.ca
Areas of Expertise*: Machine Learning, Dynamical Systems

Manuscript submission information:

Manuscript Submission Deadline: 28 February 2026

Manuscripts should be submitted using the online submission system of the journal at https://www.editorialmanager.com/PHYSD/default.aspx. Authors should select the article type ‘VSI: MLDSAIT ’ during the submission. The author guidance can be found at https://www.elsevier.com/journals/physica-d-nonlinear-phenomena/0167-2789/guide-for-authors.

Keywords:

    Machine Learning, 
    Dynamical Systems, 
    Algorithmic Information Theory