Complexity of AI

Dear Colleagues,

Modern artificial intelligence (AI) models, specifically artificial neural networks, are typically complex systems that consist of a large number of interactions. With the accelerating progress made in the capabilities of AI, there is a growing need for a better understanding of their working principles. Complexity science, the field underlying Giorgio Parisi’s work which won the 2021 Nobel prize in physics and providing some of the key theoretical foundations for the AI contributions presented by John J. Hopfield and Geoffrey Hinton, which won the 2024 Nobel prize, is receiving increasing attention in relation to the pursuit of an understanding of AI models.

As complex systems, AI models typically exhibit the phenomena of phase transition, emergence (of intelligence), multiple meta-stabilities, chaos, self-organization, etc. Studying and analyzing AI models from the perspective of complexity science can pave the way for a deeper understanding of how AI models process information, carry out logic reasoning, and generate new data like images or text. Such findings can potentially help us to improve the capabilities, safety, and efficiency of AI models through untangling their complexities.

This Special Issue is organized in conjunction with the Focused Session entitled “Complexity of AI” at the Asia-Pacific Summer School and Conference on Networks and Complex Systems (APCNCS) 2026, held in Singapore (https://apcncs2026.github.io/). Authors are invited to contribute to both the Special Issue and the Focused Session. We welcome original research exploring the complexity of AI models from diverse perspectives, including methods, theories, applications, and empirical studies.

You may choose our Joint Special Issue in Entropy.

Dr. Siew Ann Cheong
Dr. Ling Feng
Dr. Lock Yue Chew
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Complexities is an international peer-reviewed open access quarterly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1000 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• complexity science
• artificial intelligence
• neural networks
• statistical physics
• phase transitions
• dynamical systems
• self-organization