My dear AGI friends, OpenCog? NARS?AIXI? ... Which one is more intelligent? How can you measure the intelligence quotient of AGI systems? This is the question.
I am working on my new paper to be submitted to AGI-22 before the deadline 14 March (the deadline is extended thanks to the AGI-22 organizers). It is the continuation of my AGI-21 paper which was about defining a versatility index for AGI systems based on the Legg-Hutter statement. In that paper I claimed that versatility is a major necessary condition for AGI agents, and turned that statement into a formula called Versatility Index (VI). But versatility is not enough. An AGI system must satisfy some other conditions too. The new paper is the first steps towards defining a computational formula for measuring the intelligence level of AGI systems, like what IQ tests do in order to measure the intelligence quotient of humans. I call it Versatility-Efficiency Index (VEI) and it is based on both Legg-Hutter and Goertzel statement. Also, a framework for determining the complexity of different sub-spaces of the universal problem space is proposed. The formula and the framework is ready and I have written about more than half of the paper. For this paper, I aim to enrich it with comments and contributions of interested AGI researchers as my co-authors. Since the deadline is very close, I invite interested AGI researchers (preferably native English speakers) to contact me as soon as possible. Email: [email protected] Best regards, Mohammadreza Alidoust ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T73f43e502fded7a6-Me207e7b124b7dcb449a9eebc Delivery options: https://agi.topicbox.com/groups/agi/subscription
