Expected value of perfect information (EVPI) is a concept used in decision theory to measure the cost of uncertainty when making decisions. It is commonly used in health economics to determine the price one would pay for perfect information. EVPI analysis tries to measure the expected cost of that uncertainty, which can be interpreted as the expected value of perfect information.

UC Berkeley's CS 188 course covers the basic ideas and techniques for designing intelligent computer systems, emphasizing statistical and decision-theoretic modeling. By the course's end, students will have built autonomous agents that can make efficient decisions in a variety of settings.