Video

Unknown · 0:00

This talk explores why we still lack a working mathematical model of "circuits" (schemas) inside neural networks, using the successful case of sparse autoencoders (a model for representations) as a template, then exte...

Read the full summary on tuber