Expert-AI hybrid maps how urban resilience indicators depend on each other

Researchers have built a hybrid framework that combines expert knowledge with transformer-based semantic inference to map how urban resilience indicators depend on one another, according to a study published in Applied Network Science on 18 September 2026. The work, by Mozhgan PourmoradNasseri and Amir Albadvi, addresses a persistent gap between qualitative policy constructs and quantitative system models in resilience governance.

The team synthesised 451 measurable sub-indicators drawn from eleven global frameworks into the Urban System Abstraction Hierarchy, a reproducible graph dataset of 40 indicators across seven socio-economic domains. The authors note that this spans institutional, economic and social systems where earlier implementations stop at physical ones.

Methodologically, sentence-transformer embeddings project each indicator into a high-dimensional semantic space. These embeddings are combined with an expert-derived prior, so that only dependencies supported by both semantic and operational evidence are retained, yielding a directed dependency graph among indicators.

Applying the pipeline to a Vancouver case study, the researchers used complex network analysis and report a small-world topology, a modular structure in which governance-related indicators form a tightly coupled core, and a hub-sensitive robustness profile. These are model-based characterisations of the constructed network for that case, not measurements of Vancouver’s real-world resilience outcomes.

By combining directional and positional centrality measures, the analysis assigns each indicator a functional role: anchor, bridge, peripheral driver, stabilizer or receiver. This distinction separates indicators that drive downstream domains from those that accumulate system state, offering a structured way to interpret positions within the dependency network.

The authors state that the pipeline produces reproducible, internally consistent system representations that are transferable across cities and applicable to downstream policy analysis and network-based resilience modelling. The article is published open access, and the version shared is a peer-reviewed, accepted manuscript carrying a permanent DOI, subject to further edits before the final Version of Record.

Subject of Research: A hybrid expert-AI pipeline for constructing and characterising urban resilience indicator dependency networks

Article Title: A hybrid expert-AI pipeline for constructing and characterising urban resilience indicator dependency networks

Article References: PourmoradNasseri, M., & Albadvi, A. (2026). A hybrid expert-AI pipeline for constructing and characterising urban resilience indicator dependency networks. Applied Network Science. https://doi.org/10.1007/s41109-026-00831-1

Image Credits: AI Generated

DOI: 10.1007/s41109-026-00831-1

Keywords: urban resilience, dependency networks, complex network analysis, sentence transformers, indicators, Vancouver, governance, small-world topology, expert knowledge, semantic inference, hybrid, expert-AI