up vote 3 down vote favorite

Currently, we have tags , , and . Is there a difference? If so, somebody should edit the tag wikis to make it clear.

share
1  
Certainly, the wiki of lg.learning suggests that it encompasses machine-learning. – Dave Clarke♦ Feb 23 '11 at 16:57
I noticed that. There are indeed a lot of questions with both tags. I see now there was a previous meta question discussing this, so maybe we don't have to do anything. – Peter Shor Feb 23 '11 at 18:59
As a side note, my personal opinion is that it is better to leave decisions about tag merges/synonyms to the researchers in the related area since they are the people who really use those tags (unless the tags start to cause problems, e.g. repeated mistagging). – Kaveh♦ Feb 24 '11 at 4:59

2 Answers

up vote 1 down vote

I think the consensus from that meta question was that lg.learning refers to learning-theory (defined as COLT/ALT type work) and machine learning is the broader category.

share
So machine learning is actually the top-level tag in this case? – Peter Shor Feb 24 '11 at 11:11
yes, that is correct. – Suresh Venkat♦ Feb 24 '11 at 16:51
I'm not sure that things should be this way. I think of the machine-learning tag as the theory of machine learning, which is different from lg.learning. The former tag, for example, isn't concerned with learning monotone DNF, while the latter tag is. I view machine-learning tag as ICML theory track and the lg.learning tag as COLT/ALT. – Lev Reyzin Feb 25 '11 at 21:05
so lg.learning would not contain (for example) algorithms for variational inference, or graphical model work, or MCMC ? – Suresh Venkat♦ Feb 25 '11 at 22:01
right - at least in my worldview. of course, we need to all agree on the proper uses for these tags... – Lev Reyzin Feb 25 '11 at 22:28
I don't think the arxiv agrees with you though: look at this list: arxiv.org/list/cs.LG/11 – Suresh Venkat♦ Feb 25 '11 at 22:53
You're right, but I wasn't so happy when lg.learning and learning-theory became synonyms (afterwards I thought of lg.learning as learning-theory). At this point it looks like lg.learning isn't restricted to COLT/ALT stuff, so what's there to distinguish it from machine-learning? Maybe at this rate would should just make all the learning tags synonyms? – Lev Reyzin Feb 25 '11 at 23:45
1  
you're the expert :) - make a proposal – Suresh Venkat♦ Feb 26 '11 at 3:46
before I do that, do you know if it's possible to undo making two tags synonyms? – Lev Reyzin Feb 26 '11 at 18:06
I'm not sure about that. It does seem like I can delete an association though. – Suresh Venkat♦ Feb 27 '11 at 4:10
up vote -3 down vote

Although I am far from being an expert in this, I believe that lg.learning and machine-learning refer to the same thing. I do not know of any other kind of learning related to TCS other than machine learning. However, online-learning is a subtype of machine learning and I believe that it should be a different tag.

share

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged