Parallel Online Learning
Jan 19, 2010 6282 views 
en-de
en-es
en-fr
en-sl
en
en-zh
0.25
0.5
0.75
1.25
1.5
1.75
2
A fundamental limit on the speed of training and prediction is imposed by bandwidth: there is a finite amount of data that a computer can access in a fixed amount of time. Somewhat surprisingly, we can build an online learning algorithm fully capable of hitting this limit. I will discuss approaches for breaking the bandwidth limit, including empirical results.