An Information-Geometric Justification for Composite Coherence in Event-Based Narrative Extraction
arXiv:2606.29118v1 Announce Type: cross Abstract: Graph-based narrative extraction relies on a coherence function to score transitions between events, but the coherence metrics in current use are defined operationally and lack an information-theoretic foundation. We study the composite metric $C=\sqrt{A\cdot T}$, where $A$ is the angular similarity of document embeddings and $T=1-d_{\mathrm{JS}}$ is a topic proximity from the Jensen-Shannon distance of soft memberships, and give it an information-geometric reading together with an axiomatic characterization of the geometric-mean combinator. On
This research provides a more robust theoretical foundation for event-based narrative extraction, which is crucial as the complexity and volume of information for AI systems continue to grow.
A more principled approach to understanding and extracting narratives from vast datasets will improve the reliability and interpretability of AI agents and information retrieval systems.
The operational definition of coherence in graph-based narrative extraction is updated with an information-theoretic and geometric justification, potentially leading to more accurate and robust narrative analysis.
- · AI developers
- · Information retrieval systems
- · Natural language processing researchers
- · Less robust AI models
- · Information overload
- · Manual narrative analysis
Improved coherence metrics will lead to more effective narrative extraction from complex data.
Better narrative understanding could enhance the capabilities of autonomous AI agents in interpreting and responding to real-world events.
More sophisticated narrative extraction may empower new forms of automated analysis in social sciences, market intelligence, and geopolitical forecasting.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG