BLUE
ML
Malcolm Lett
@malcolml.bsky.social
Interested in the computational basis of consciousness and intelligence. malcolmlett.github.io/ medium.com/@malcolmlett
22 followers27 following64 posts
MLmalcolml.bsky.social

Results are from training a simple binary classifier against an SKLearn "circles" dataset. Three layers: 10-unit relu, 5-unit relu, 1-unit sigmoid.

0
MLmalcolml.bsky.social

Running some experiments to see how the fraction of active units in a network affects learning rates (ie: have a non-zero output activation), and the first result is very promising - looks like a clear indication that low activity rates badly affects learning rates. #MLSky

Loss curve compared against fraction of active units in each network layer
1
MLmalcolml.bsky.social

Surprising outcome #4: consciousness is an emergent phenomenon, that depends on having the right kind of structural organisation

0
MLmalcolml.bsky.social

Surprising outcome #3: our conscious perception of our thoughts may be only a tiny fraction of our actual thoughts.

0
MLmalcolml.bsky.social

Surprising outcome #2: consciousness is "post-causal": having no influence over the event to which we are conscious, but having direct influence over subsequent events.

0
MLmalcolml.bsky.social

Surprising outcome #1: the cartesian theatre metaphor is actually a very good metaphor for conscious function.

0
ML
Malcolm Lett
@malcolml.bsky.social
Interested in the computational basis of consciousness and intelligence. malcolmlett.github.io/ medium.com/@malcolmlett
22 followers27 following64 posts