Reproduce “A Measure of Synergy based on Union Information” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
André F.C. Gomes, Mário A.T. Figueiredo
Researchers from the University of Lisbon introduced I_CI_∪, a new measure for union information, and its derived synergy measure S_CI within the Partial Information Decomposition framework, quantifying synergy as information beyond conditional independence. This measure satisfies extended Williams-Beer axioms, avoids problematic negative synergy, and is computationally efficient while aligning with established values for benchmark distributions like XOR.
Key insights
- Synergy can be robustly defined and interpreted as the portion of total information about a target that cannot be captured by assuming conditional independence among the sources.
- A critical review of existing PID properties identifies a consistent set of axioms for union information measures, distinguishing between robust properties and those that may impose arbitrary constraints.
- The proposed measure offers computational simplicity by avoiding complex optimization problems, making it more accessible and scalable for analyzing real-world complex systems.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.