Training iterations

[stiangithub]

So the simulation's training iteration is 0? So how does the model be trained? And why does the plot show that regret is decreasing?

[huazhengwang]

Training iterations refer to offline training, where no regret is evaluated. Normally an online/bandit algorithm only have online training & testing (refer to testing iterations). We add training iterations to support offline warm-start, and it is generally set to be 0 in all ours and baseline algorithms.

What do you mean by regret is decreasing?