Certainty vs. Intelligence скачать в хорошем качестве

Certainty vs. Intelligence

In a keynote talk for the AISoLA conference held in Crete in Fall 2024, Edward Lee from Berkeley postulates that the certainty that rigorous engineering methods traditionally yield may be fundamentally incompatible with integrating intelligence into engineered systems. We may be forced to choose whether robots, self-driving cars, and other autonomous systems are intelligent or demonstrably safe. He makes this provocative case two ways, one using the theory of Bayesian learning and a second more philosophical argument that questions whether intelligence is based on rationality. If this contradiction is fundamental, then the question becomes, how do we get the advantages of intelligence with acceptable risk?