Pei Wang, Temple University, USA (in person)
While the term “Artificial General Intelligence” (AGI) has been used to identify a research field for 20 years, a consensus on its goal remains controversial. These divergent interpretations are not merely philosophical, as they lead to fundamentally different research agendas and engineering methodologies. In this talk, the NARS (Non-Axiomatic Reasoning System) project is introduced, with an outline of its assumptions, representations, inference rules, and working process. NARS offers a fundamentally different vision of AGI than the one implied by today’s LLMs.