open-connectome-classes / StatConn-Spring-2015-Info

introductory material
18 stars 4 forks source link

Recurrent vs all-to-all connectivity #37

Open whock opened 9 years ago

whock commented 9 years ago

In the neuroscience literature, the idea of recurrent networks are important as that architecture is thought to have important properties. One example is being able to instantiate content-addressable memory (Hopfield 1982, Hopfield and Tank, 1986) which is thought to occur in the recurrent CA3 region of the hippocampus. But when they talk of these nets I usually don't see them using the term all-to-all despite them having apparently the same connectivity matrix.

My question is that it seems like recurrent networks and all-to-all networks have the same connectivity matrix, is this right? Or is there a difference I'm missing? And if they're the same why does the neurosci literature talk a lot about recurrent nets and not all-to-all nets if they're same wiring?

jovo commented 9 years ago

weights. recurrent networks are all about setting the weights properly.

On Tuesday, February 3, 2015, whock notifications@github.com wrote:

In the neuroscience literature, the idea of recurrent networks are important as that architecture is thought to have important properties. One example is being able to instantiate content-addressable memory (Hopfield 1982, Hopfield and Tank, 1986) which is thought to occur in the recurrent CA3 region of the hippocampus. But when they talk of these nets I usually don't see them using the term all-to-all despite them having apparently the same connectivity matrix.

My question is that it seems like recurrent networks and all-to-all networks have the same connectivity matrix, is this right? Or is there a difference I'm missing? And if they're the same why does the neurosci literature talk a lot about recurrent nets and not all-to-all nets if they're same wiring?

— Reply to this email directly or view it on GitHub https://github.com/Statistical-Connectomics-Sp15/intro/issues/37.

the glass is all full: half water, half air. openconnecto.me, we're hiring! https://docs.google.com/document/d/14SApYAzxF0Ddqg2ZCEwjmz3ht2TDhDmxyZI2ZP82_0U/edit?usp=sharing , jovo.me, my calendar https://www.google.com/calendar/embed?src=joshuav%40gmail.com&ctz=America/New_York

whock commented 9 years ago

great thanks!