One of the duties of social Intelligence is to create consensus on symbols which prove meaningful for the pursuit, description and achievement of objectives shared by most agents.
[ quorum strategies for clustering performed by several agents, object discovery, object permanence, aesthetic]
A subsequent duty of intelligence is to then efficiently transform senses, sensors and signals into said symbols, so as to enclose and embed as much of the original saliency of the accessible information.
[feature extraction, dimensional reduction, semantic data compression]
Abstract intelligence is responsible for organising symbol streams into data structures to as to facilitate the more likely fulfilment of reward patterns or defined functions.
[ symbolic solvers, wolfram alpha]
Emotional intelligence is the module that helps the agents manage, weigh and choose actions or responses based on the observed symbol and signal states.
[ multi-objective optimisation ]
Emotional intelligence also helps us simulate and predict hidden reactions and outcomes of other agents under known or assumed conditions.
[ collaborative reinforcement learning under uncertainty and imperfect, asymmetrical communication ]
Introspection reprocesses existing data using alternative models, so as to achieve better embeddings, latent spaces which are more relevant for desired purpose, stronger compression or faster processing.
[epistemology, neural data structures, meta-models]
This post first appeared on Bogdan Bocse - Solutions Architecture, Business And Random Thoughts, please read the originial post: here
