2011 · Artificial intelligence
Probabilistic and causal models made different kinds of questions explicit.
Visual guide
Describe statistical relationships.
State causal assumptions.
Ask what a deliberate change would do.
Causal reasoning provides a structured language for analysing interventions, explanations and counterfactual questions.
Causal conclusions require assumptions and suitable evidence; a correlation graph alone cannot establish them.
01 · The idea
Judea Pearl developed mathematical tools for reasoning under uncertainty and about causation. A probabilistic graph can represent dependencies; a causal model adds assumptions about how changes propagate. This distinction lets us separate observing that two things occur together from predicting what an intervention would do.
Causal reasoning provides a structured language for analysing interventions, explanations and counterfactual questions.
02 · The people
Judea Pearl developed frameworks for reasoning under uncertainty and for distinguishing association from causation. His work on Bayesian networks organized probabilistic dependencies through graphs, enabling efficient inference in complex models. He later developed graphical methods for causal reasoning and interventions. At UCLA, these ideas connected artificial intelligence with statistics and scientific explanation, helping researchers ask what would happen if a system were deliberately changed.
Read the biography ↗03 · Read further
Original explanations based on the official award record. The subject label is an editorial navigation aid.
Reinforcement learning connects actions, rewards and improving predictions.
The story
Deep networks learn useful representations through many layers of computation.
The story