Hotz, Peter Eggenberger, Luchsinger, Rolf, Weyland, Mathias, Enayati, Aref, Jamieson, W. David  ORCID: https://orcid.org/0000-0001-8260-5211, Castell, Oliver ORCID: https://orcid.org/0000-0002-6059-8062 and Füchslin, Rudolf M.
2026.
Emergent predictability from parameter ambiguity in biochemical networks.
Presented at: ALIFE 2026,
Ontario, Canada,
17-21 August 2026.
ALIFE 2026: Proceedings of the 2026 Artificial Life Conference.
ASME,
10.1162/ISAL.a.999
|
Preview |
PDF
- Published Version
Available under License Creative Commons Attribution. Download (1MB) | Preview |
Abstract
Complex adaptive systems often involve many interacting components and parameters, leading to strong non-identifiability: distinct parameter combinations can produce indistinguishable system behavior. Optimization methods may therefore identify parameter sets that fit observed data equally well yet differ substantially in their individual values. Despite this microscopic variability, the resulting models often retain comparable predictive power. We investigate this phenomenon in nonlinear biochemical reaction networks and show that predictive robustness arises from a low-dimensional structure in parameter space. Using Fisher Information Matrix (FIM) analysis, we identify a dominant collective parameter, Λ1, that acts as an effective scale controlling the system’s temporal and amplitude response. Perturbations that preserve Λ1 leave model predictions largely unchanged, whereas perturbations that alter Λ1 substantially degrade predictive accuracy. This indicates that system behavior is governed by a low-dimensional structure embedded in a high-dimensional parameter space. These findings provide a mechanistic explanation for how robust, reproducible system-level behavior can emerge from parameter ambiguity in complex adaptive systems.
| Item Type: | Conference or Workshop Item - published (Paper) |
|---|---|
| Date Type: | Published Online |
| Status: | Published |
| Schools: | Schools > Pharmacy |
| Publisher: | ASME |
| Date of First Compliant Deposit: | 30 September 2026 |
| Last Modified: | 30 Sep 2026 11:36 |
| URI: | https://orca.cardiff.ac.uk/id/eprint/189909 |
Actions (repository staff only)
 |
Edit Item |
Dimensions
Dimensions