Epstein Secrets

Some Al theorists would argue that this division into "feature identification" versus "action selection" is unnecessarily artificial; for instance, Hawkins 11113011 or Arel [ARC09bl might sug- gest to use a single hierarchical neural network to do both of them. But the brain after all t in game theory,

LN hierarchies and explicitly constituting an intermediate level of representation between attractors in DeSTIN and the habitual cognitive usage of CogPrime Atoms and Atom-networks. For simple reference we refer to this as the "Semantic CSDLN" approach. A "tripartite semantic CSDLN" consisting of interlinked

, and suggested answers to each of them ISot III. Stephen Omohundro 1Omo081 has argued that any advanced A

antic net because of the broad scope of the Atoms in the network: for example, Atoms may represent percepts, procedures, or parts of concepts. Most CogPrime Atoms have no corresponding English label. However, most CogPrime Atoms do have probabilistic truth values, allowing logical semantics. EFTA00624031 2

t, and suggested answers to each of them [Sot11]. Stephen Omohundro [Omo08] has argued that any advanced AI

emantic net because of the broad scope of the Atoms in the network: for example, Atoms may represent percepts, procedures, or parts of concepts. Most CogPrime Atoms have no corresponding English label. However, most CogPrime Atoms do have probabilistic truth values, allowing logical semantics. HOUSE_OVERSIGHT_0