Visual Learning and Recognition: Difference between revisions

No edit summary
Line 11: Line 11:
If we have very few images, we are working on an extrapolation problem.   
If we have very few images, we are working on an extrapolation problem.   
As we approach an infinite number of training samples, learning becomes an interpolation problem.   
As we approach an infinite number of training samples, learning becomes an interpolation problem.   
Traditional datasets are in the order of \(10^2-10^4\) training samples.   
Traditional datasets are in the order of <math>10^2-10^4</math> training samples.   
Current datasets are in the order of \(10^5-10^7\) training samples.
Current datasets are in the order of <math>10^5-10^7</math> training samples.


In tiny images <ref name="torralba2008tinyimages"></ref>, Torrabla ''et al.'' use 80 million tiny images.
In tiny images <ref name="torralba2008tinyimages"></ref>, Torrabla ''et al.'' use 80 million tiny images.