Document worth reading: “Towards a framework for the evolution of artificial general intelligence”
In this work, a novel framework for the emergence of general intelligence is proposed, the place brokers evolve by approach of environmental rewards and be taught all via their lifetime with out supervision, i.e., self-supervised learning by approach of embodiment. The chosen administration mechanism for brokers is a biologically plausible neuron model based totally on spiking neural networks. Network topologies develop to be further sophisticated by approach of evolution, i.e., the topology is simply not fixed, whereas the synaptic weights of the networks cannot be inherited, i.e., new baby brains shouldn’t educated and don’t have any innate data of the setting. What is matter to the evolutionary course of is the neighborhood topology, the kind of neurons, and the kind of learning. This course of ensures that controllers that are handed by approach of the generations have the intrinsic potential to be taught and adapt all through their lifetime in mutable environments. We envision that the described technique would possibly end in the emergence of the best form of artificial general intelligence. Towards a framework for the evolution of artificial general intelligence