6.S191
Image Domain Transfer: enabling machines to have human-like imagination abilities
http://introtodeeplearning.com/2019/materials/2019_6S191_L9.pdf
MUNIT assumption: Partially Shared LatentSpace
C: content space
S1, S2 -> style space
X1, X2 : domain
encoding
- X1 -> C, X1 -> S1
- X2 -> C, X2 -> S2
translation
- X1 -> C-> S1-> X2
network architecture
Supervised vs Unsupervised
supervised
{xi, yi}, i : N
unsupervised
{xij, yij} , i, j : N
Biologically Inspired Neural Networks (IBM)
http://introtodeeplearning.com/2019/materials/2019_6S191_L8.pdf
https://github.com/DimaKrotov/Biological_Learning
Update of weignts
- The update of backpropagation learning is non-local.
- The update of biological learning is local.
Hebbian learning
Organization of Behavior (Hebbian theory)
http://s-f-walker.org.uk/pubsebooks/pdfs/The_Organization_of_Behavior-Donald_O._Hebb.pdf
Synaptic plasticity rule
https://scholar.google.co.jp/scholar?q=Synaptic+plasticity+rule&hl=ja&as_sdt=0&as_vis=1&oi=scholart
Alternative ideas on biologically plausible learning
[1801.00062] Dendritic error backpropagation in deep cortical microcircuits
Lp space (Lebesgue norm p)
https://en.wikipedia.org/wiki/Lp_space
etc
Mathematics Stack Exchange
continuing ...