Lecture – Probabilistic Reasoning and Reasoning Under Uncertainty - 5DV244 Human-Centered Artificial Intelligence 2026

Date: 2026-09-17

Location: Hörsal NAT.D.470

Teacher: Esteban Guerrero (EG)

Week: 38

Many real-world AI applications operate under uncertainty, where information may be incomplete, ambiguous, noisy, or probabilistic. This lecture introduces probabilistic reasoning and methods for making decisions when certainty cannot be guaranteed.

We move from deterministic reasoning to reasoning with degrees of belief, exploring how probabilistic models can represent uncertain knowledge and support intelligent decision-making.

The lecture also connects probabilistic reasoning with Bayesian networks and decision theory, showing how an intelligent agent can use uncertain information to choose appropriate actions.

• Reasoning under uncertainty
• Probabilistic reasoning
• Probability and degrees of belief
• Bayesian reasoning and inference
• Bayesian networks
• Causal reasoning and interventions
• Decision-making under uncertainty
• Expected utility and rational action

• Understand why uncertainty is unavoidable in many AI applications.
• Explain the difference between logical reasoning and probabilistic reasoning.
• Use probabilistic models to represent uncertain knowledge.
• Understand the basic principles of Bayesian inference.
• Understand how Bayesian networks represent dependencies between variables.
• Understand the role of interventions in causal reasoning.
• Understand how AI systems make decisions when information is incomplete.
• Explain the relationship between probability, utility, and rational decision-making.