Marco, N. D., Cinelli, M., Nakano, S., & Frosini, A. (2026). Optimal and heuristic strategies for evaluating the influence of coordinated behavior in information cascades and retweet networks. arXiv [cs.SI].
Summary
This paper tackles a question that it argues has been underexplored in research on Coordinated Inauthentic Behavior (CIB): not whether coordinated accounts can be detected, but how much they actually shape information diffusion once identified. The authors contend that the field has concentrated heavily on detection at the expense of quantifying downstream impact. To address this, they propose two complementary post-hoc evaluation frameworks — one built on information cascades, the other on retweet networks — and develop both optimal and heuristic strategies for measuring the influence of coordinated accounts within these diffusion structures.
Key Contributions
- Introduces two complementary frameworks for the post-hoc evaluation of coordinated accounts’ influence on information diffusion.
- Explicitly reframes CIB research away from detection and toward influence quantification.
- Provides both optimal and heuristic strategies, making the approach applicable across information cascades and retweet networks.
Methods
The authors formalize the CIB influence-evaluation problem over information cascades, and introduce a second, complementary formulation grounded in retweet networks. For each setting they derive optimal strategies for assessing how much coordinated accounts contribute to diffusion outcomes, and pair these with heuristic alternatives — presumably to retain tractability where exact optimization is costly. The work is primarily methodological/algorithmic in character rather than empirical.
Findings
- Both optimal and heuristic strategies are presented as viable for assessing coordinated-account influence in diffusion structures.
- Specific empirical results are not available from the provided summary, which cuts off before describing experimental outcomes.
Connections
This paper sits at the intersection of CIB detection and network-structure analysis, extending a largely detection-focused literature toward measuring downstream diffusion impact; it connects naturally to detection work that this paper positions itself against, such as Nizzoli2020-cf and Luceri2025-tr. Its focus on retweet networks and cascade structure links it to structural coordination studies like Minici2024-tf, Giglietto2020-9d8acdd7, and Keller2019-nk, while sharing an author lineage and thematic concern with Di-Marco2025-aa.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify