Important Dates

  • Submission deadline: March 10, 2026 at 11:59 PM AOE
  • Notification of acceptance: March 13, 2026
  • Early-bird registration deadline: March 20, 2026
  • Satellite date: June 1, 2026

Organizers

  • Sahil Loomba, Chapman–Schmidt AI in Science Fellow, Department of Mathematics, Imperial College London
  • Dean Eckles, William F. Pounds Professor of Management, MIT Sloan School of Management
  • Johan Ugander, Associate Professor, Department of Statistics and Data Science, Yale University

Over the past decade, the fields of network science and causal inference have each made significant statistical and algorithmic advances. Understanding node-to-node connections and cause-and-effect relationships are each seen as fundamental to studying complex natural and social systems. There is a natural synergy between these two perspectives: connections between nodes can act as conduits of causal influence, while external interventions can selectively precipitate node-level connections. However, much of this progress has occurred independently. The network science community often studies causal relationships in the context of parametric network spreading processes that generate dynamic network data, whereas the causal inference community typically addresses the non-parametric case of a single snapshot network observation leading to fundamental non-identifiability. The former approach can be more efficient and interpretable, while the latter is potentially more robust to model misspecification. In the spirit of this “bias–variance trade-off,” we argue that there are substantial missed opportunities at the intersection of these two approaches — and indeed, their respective communities — that could benefit from cross-fertilization of methods, problems, and perspectives.

Aims

To this end, the inaugural Causal Network Science (CausNetS) satellite has the overall objective to advance the field of network science toward a causal network science. The satellite will:

  • Introduce researchers in network science to causal inference methods that are particularly network-relevant;
    • Showcase cutting-edge work on experiment design, causal identification, and estimation in networks.
    • Identify methodological gaps and grand challenges at the intersection of networks and causality.
  • Foster a sustained community of theoretical and applied network scientists thinking about causal questions;
    • Develop a shared language and set of conceptual frameworks for causal reasoning on networks.
    • Connect methodological advances with problems in epidemiology, online platforms, field experiments.
  • Establish Causal Network Science as an emerging subfield with active scholarly exchange.
    • Create a working repository of problems, datasets, and examples for causal networks research.
    • Outline plans for collaborations, future satellite meetings, and cross-community initiatives.

Expected Outcomes

By the end of the satellite, participants will have a clear and strong sense of Causal Network Science as a significant emerging subfield, essential for answering causal questions in complex natural and social systems. Invited speakers will equip network scientists with core concepts and challenges from causal inference, and contributed talks will showcase the research frontier on experimentation, identification, and estimation under network interference. By stimulating new collaborations and outlining plans for future initiatives, CausNetS aims to establish itself as a recurring event with a growing, interdisciplinary community.