Journal article · 2026 · Open access
Fractal basin geometry and the limits to predictability and reproducibility in deep learning
Nature Machine Intelligence · Published 5 November 2026 · DOI 10.1038/s42256-026-01336-3
In brief
“Fractal basin geometry and the limits to predictability and reproducibility in deep learning” is a 2026 journal article by Andrew Ly and Pulin Gong, published in Nature Machine Intelligence. OpenAlex counts 1 citations as of 16 November 2026, which puts it in the top 9% of works published in 2026 by citations. It is open access under a CC BY licence: the full text is free to read at Nature Machine Intelligence.
- Citations
- 1
- Most cited in
- 2026 (1)
- Field-weighted impact
- 1.40
- References
- 55
- Authors
- 2
Abstract
Fundamental limits to predictability are central to our understanding of many physical and computational systems. In deep learning, training outcomes are often difficult to predict and reproduce even when extrinsic sources of randomness are controlled, yet the underlying mechanism remains unclear. Here we propose riddled basin geometry as a unifying dynamical systems mechanism: every initialization leading to one solution lies arbitrarily close to initializations leading to different solutions, forming a fat fractal with structure at arbitrarily fine scales. Using a minimal neural network model, we analytically establish the classical route to riddling and derive a near-zero uncertainty exponent: even sharp increases in initialization precision yield only marginal gains in outcome predictability. In deep networks, intertwined training destinations and near-zero parameter-space exponents provide robust signatures consistent with riddling across the tested architectures, optimizers and training protocols, offering a common dynamical account of otherwise disparate empirical observations. We further introduce functional uncertainty exponents, extending these diagnostics of scale-insensitive irreproducibility from parameter space to the learned functions themselves. Together, our results support riddled basin geometry as a candidate organizing principle for irreproducibility in deep learning, with practical implications for optimization and the reliable deployment of artificial intelligence.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Ly, A., & Gong, P. (2026). Fractal basin geometry and the limits to predictability and reproducibility in deep learning. Nature Machine Intelligence. https://doi.org/10.1038/s42256-026-01336-3
MLA 9
Ly, Andrew, and Pulin Gong. “Fractal basin geometry and the limits to predictability and reproducibility in deep learning.” Nature Machine Intelligence, 2026. https://doi.org/10.1038/s42256-026-01336-3.
Chicago
Ly, Andrew and Pulin Gong. 2026. “Fractal basin geometry and the limits to predictability and reproducibility in deep learning.” Nature Machine Intelligence. https://doi.org/10.1038/s42256-026-01336-3.
Harvard
Ly, A. and Gong, P. (2026) ‘Fractal basin geometry and the limits to predictability and reproducibility in deep learning’, Nature Machine Intelligence. doi:10.1038/s42256-026-01336-3.
IEEE
A. Ly and P. Gong, “Fractal basin geometry and the limits to predictability and reproducibility in deep learning,” Nature Machine Intelligence, 2026, doi: 10.1038/s42256-026-01336-3.
BibTeX
@article{ly2026fractal,
title = {{Fractal basin geometry and the limits to predictability and reproducibility in deep learning}},
author = {Ly, Andrew and Gong, Pulin},
journal = {Nature Machine Intelligence},
year = {2026},
doi = {10.1038/s42256-026-01336-3},
url = {https://doi.org/10.1038/s42256-026-01336-3}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
- outcome predictability
- initialization sensitivity
- dynamical systems
- neural network training
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Ly, A., & Gong, P. (2026). Fractal basin geometry and the limits to predictability and reproducibility in deep learning. Nature Machine Intelligence. https://doi.org/10.1038/s42256-026-01336-3
How many times has this paper been cited?
1 times, according to OpenAlex on 16 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at Nature Machine Intelligence: sign in and open the PDF from this page.
Who wrote this paper?
Andrew Ly, Pulin Gong.
What is the DOI of this paper?
10.1038/s42256-026-01336-3, permanently at https://doi.org/10.1038/s42256-026-01336-3. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 16 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.