Wednesday, November 18, 2026

2:30 PM - 3:30 PM

Sutherland Theatre

---

Presentation Outline

The evolution toward the Internet of Battlespace Things (IoBT) and digital twins has made machine learning indispensable for tactical decision-making. However, traditional statistical AI remains inherently brittle in contested domains. Because correlation-based models rely on historical patterns, they struggle when unexpected conditions or novel adversary tactics disrupt operations. Resilient network defense requires AI to graduate from pattern recognition to robust causal reasoning. This talk introduces causal learning as a foundational capability for military communications. An accessible primer on core causal principles illustrates how establishing true cause-and-effect relationships overcomes the vulnerabilities of statistical AI, delivering adaptable and explainable decision support. The broad applicability of causal modeling is demonstrated through two critical operational focus areas. First, historical tactical network telemetry suffers from severe confounding bias because compensatory routing protocols continuously adapt to unobserved conditions. Causal modeling systematically untangles these hidden biases to uncover the true impact of network interventions. Second, causal frameworks empower network digital twins to safely evaluate counterfactual scenarios. This shifts anomaly mitigation from a fragile, reactive posture to a predictive capability that anticipates systemic impacts before operators execute high-risk interventions. Ultimately, this talk provides a pragmatic roadmap for integrating causally-aware decision support into next-generation Command and Control systems.

---

Speaker