Jose Hernandez-Orallo (Valencia/Cambridge), “The Generality of Natural and Artificial Intelligence: Task Difficulty as the Elephant in the Room”

Abstract: Understanding and recreating intelligence is possibly the biggest scientific challenge of our time. Evolution has produced organisms that are highly specialised for some cognitive tasks, whereas others present what has been called general intelligence, with humans identified as the paragon. Artificial intelligence (AI), despite decades of efforts to achieve generality, is still specialised. It is a major research question to disentangle the notion of general intelligence, by clearly determining what generality is and how it can be measured for individuals rather than populations. Under limited resources, we must overhaul the classical yet misleading interpretation of general intelligence as ‘success in all sorts of situations’ and introduce a new view of generality as ‘comprehensive performance up to a level of difficulty’. The degree of generality then refers to the way an agent’s capability is distributed as a function of task difficulty, according to environmental and cognitive pressures. This dissects the notion of general intelligence into two non-populational measures, generality and capability. We interpret and apply these measures with humans, non-human animals and AI systems. The choice of the difficulty function now plays a prominent role in this new conception of generality, which brings a quantitative tool for shedding light on long-standing questions about the evolution of general intelligence and the evaluation of progress in Artificial General Intelligence.

Hosted by Dimitri Coelho Mollo

The Zoom Link will be sent the day before the lecture. (Contact communication@scioi.de for specific questions)