1
$\begingroup$

In my head, these two words (learning and training) seem to somehow have a fuzzy boundary between them.

For example, the word learning for me conveys the idea of training; if I want to learn something, I train myself in it!

When talking about machine learning algorithms, this happens too. I often find myself wondering if it's right to say "unsupervised training"! since, in my head, the word training is tightly related to supervised learning (here too can I say supervised training!).

Is there a clear boundary I can learn or should I just train myself the words' uses as they are :-)

$\endgroup$

1 Answer 1

0
$\begingroup$

Learning comprises the whole process, which includes a training step. There are another steps that differ from the initial learning, for example, learning reinforcement, when you re-train the model with new observations and data preprocessing.

To give you an example, lets say you have a dataset and you r model is a linear regression, the goal is to find some constants a and b that give you the best estimate of some observed outcomes. To train this model you can use different approaches: gradient descend, the explicit formula, genetic algorithms, you can train it using cross validation, using the complete dataset, etc.

$\endgroup$
4
  • $\begingroup$ Thank you for the answer but I don't think you understood my question! $\endgroup$ Commented Jul 18, 2017 at 10:38
  • $\begingroup$ Maybe I didn't, are you asking for the boundary between the concepts of training and learning? Can you please explain the question? Thanks. $\endgroup$ Commented Jul 18, 2017 at 17:41
  • $\begingroup$ Well I guess you answered part of the question by noting that the training is part of the whole process called learning. But I'm more interested in the terminology and its use, for e.g. why is the phrase "unsupervised training" rarely used even though there is a training step in unsupervised learning. $\endgroup$ Commented Jul 18, 2017 at 21:12
  • $\begingroup$ I don't know the answer to that but I guess it is because it is only used you want to be very specific. We as humans often tend to be lazy in the way we speak and assume everyone gets the meaning of the message. It is like when you say "can you hand me that thing (and point at the thing)?" instead of "can you hand me red shirt that is over the table but under the books". I think it is practical. But yes, maybe some times it could be better to be specific. Regarding unsupervised learning, I've always thought that those algorithms don't learn in the strict sense of the word but they explore. $\endgroup$ Commented Jul 19, 2017 at 1:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.