Other
Deep Learning:
For high level abstraction (learn complicated functions) we need deep architectures. It consists of multiple levels of nonlinear operation (neural nets with many hidden layers). Searching the parameter space of deep architecture is a difficult task but learning algorithms such as those for Deep Belief Networks can handle this problem with success.
High level abstraction means a man in the image whether sitting or standing. It consists of high dimensional data made up of many observed variables. We can relate small variations in geometric factors (position, orientation and lighting) with changes in pixel intensities for all pixels in an image called factors of variation. If a machine gets to know these factors and how they interact to generate the kind of data we observe.
Depth of architecture refers to levels of combination
matlab
No comment