Скачать с ютуб Hybrid computation on virtual machinery steroids: dualism to manifold discontinuities in mind-brain в хорошем качестве

Hybrid computation on virtual machinery steroids: dualism to manifold discontinuities in mind-brain

In the 1980s, connectionists claimed they could explain all brain computation in the statistical terms of neural nets. In the talk from which this is taken ("Integrating Connectionist and Symbolic Computation in Grammatical Theory"), Paul Smolensky discussed the need for hybrid models (a form of dualism). Then Luc Beaudoin of SFU and CogZest pointed out that (a) Seth Grant has shown computation happening at the post-synaptic juncture (essentially: sub-connectionist); (b) there are many layers in virtual machines; and (c) it seems the brain contains multiple manifold hierarchies from synaptic computation up to the highest levels of virtual machinery. So the arguments that connectionists originally made against symbolic modelers apply to connectionist modeling too. Smolensky concludes essentially by saying that "we need to simplify using tools until we figure out how to work with additional complexity", which is essentially what many of us have been saying to connectionists in the 1980s. The complete original talk is here:    • DCS | Paul Smolensky: Integrating Con...