Meta-learning from Learning Curves Forum

Go back to competition Back to thread list Post in this thread

> Semantics of the index in the learning curves

Hi,

the learning curves are basically two lists, one containing the accuracy (e.g. [0.5, 0.6, 0.65]) and the timestamp (e.g. [120, 300, 400]). My question is: what is the meaning of every position on these lists? Is every position one iteration of the algorithm, for instance, the iteration 1 took 120 time units, while iteration 2 took 180 time units? or there is actually no meaning in this indexing and they just correspond to ordered samples of the learning curve?

Best,

Pineda

Posted by: pineda @ Jan. 27, 2022, 12:16 p.m.

Hi,

They are algorithm iterations, but not on a linear scale. We cannot provide more details on this for this round.

Best regards,
Hung

Posted by: hungnm @ Jan. 27, 2022, 10:31 p.m.
Post in this thread