How did a mathematical framework for reasoning with uncertainty revolutionize AI in the 1980s?

The answer

Bayesian networks, formalized by Judea Pearl at UCLA in his 1988 book 'Probabilistic Reasoning in Intelligent Systems,' provide a graphical framework for representing and reasoning about uncertain knowledge. Each node in the directed acyclic graph represents a variable, and edges encode probabilistic dependencies, allowing efficient computation of how evidence about some variables changes beliefs about others via Bayes' theorem. Pearl's key insight was that the graph structure captures conditional independence relationships, dramatically reducing the number of parameters needed - from exponential to manageable. Bayesian networks became foundational across AI, powering applications from medical diagnosis (e.g., the QMR-DT system) and spam filtering to genetic analysis, and Pearl received the ACM Turing Award in 2011 for his contributions to probabilistic and causal reasoning.