| Contrastive divergence |
|
|
|
| Aim is to minimize the amount by which a
step |
||
| toward
equilibrium improves the data distribution. |
||
|
|
|
| distribution
after |
||
| one step of |
||
| Markov
chain |
||
|
|
| data |
|
| distribution |
|
|
|
| model’s |
|
| distribution |
|
|
|
|
|
| Maximize the |
|||
| divergence
between |
|||
| confabulations
and |
|||
| model’s
distribution |
|||
|
|
|
|
| Minimize
divergence |
|||
| between
data |
|||
| distribution
and |
|||
| model’s
distribution |
|||
|
|
| Minimize |
|
| Contrastive |
|
| Divergence |
|