10

Proposal: Artificial Intelligence

There is already a site about machine learning named Cross Validated. How is this proposal different?

  • @kenorb, you didn't correct the link. This discussion was about an older proposal, not the current one. You should not change posts about other proposals to look as if they are about another one. If you want a discussion about the issue you should start a new one and just link to this. – Kaveh Jun 30 '16 at 8:04
  • Related: meta.ai.stackexchange.com/q/1123/8 (access link). – kenorb Aug 10 '16 at 16:57
15

First of all, artificial intelligence is a much broader term than machine learning. While at the same time Cross Validated is not about machine learning, but about statistics (while machine learning is a field in the intersection of computer science and statistics).

  • 3
    This misconception led to several “machine learning” proposals that failed to get off the ground. In fact, Cross-validated is about statistics and machine learning, see stats.stackexchange.com/help/on-topic. – Gala Sep 8 '13 at 20:20
  • 4
    This does not contradict my answer. ML is strongly related to statistics, which does not mean that it covers whole topic as well as AI is much broader – lejlot Sep 8 '13 at 20:44
  • 1
    My point is that Cross-Validated (CV) is intended to cover machine learning fully and not only as a side-effect of its relation to statistics (again, read the help or the various discussions about this). It does therefore seem to contradict the bit about CV “[not being] about machine learning” but anyway the most important thing is to realize that CV is not merely devoted to statistics. – Gala Sep 9 '13 at 7:16
  • I am not trying to say, that what was meant is that CV is not about ML, I am rather refering to the current state. And currently there are just 1,300 question tagged as ML with 160 tagged as SVM... which shows the problem. Maybe the way out would be to better "advertise" CV as a ML site? Of course it is rather a meta.stats discussion. But there are many ways to do this, ie. adding wildcard redirect ml.stackexchange.com->stats... (For me, "stats." address already suggest "statistics"). – lejlot Sep 9 '13 at 7:29
  • 1
    CV is not a large site but I don't see how that's a problem, much less a problem that could be fixed by creating yet another site or fiddling with the URL. Getting enough good questions and experts to answer them is difficult and sites clearly labeled “machine learning” have not been more successful at it. In any case, the only thing I wanted to point out is that Cross-Validated is, among other things, about machine learning, without qualifications. You can ask ML questions, receive answers and read older questions about the topic there. (Artificial intelligence is another story of course.) – Gala Sep 9 '13 at 8:02
8

Examples of questions that are no fit for CV:

  • How does A* work?
  • Why does my genetic algorithm fail? (well, don't know whether they would accept this, as it's an optimization problem)
  • How does clause learning work in DPLL SAT-solvers?
  • Is this a real multi-agent problem?
  • Why is hierarchical planning undecidable?
  • Which description logic language should I choose for modeling X?
5

Cross-Validation: "Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts"

I somehow do not feel the relation very strongly. The only point it really is a crossover on it the word 'data'.

I will however take your word for the topic of machine learning being prominent on 'Cross-Validation', but to that I would point out, as Lejlot did before, That Machine Learning is a single aspect relating to Artificial Intelligence.

In addition to learning algorithms (On which Machine learning strongly relates) the field of AI reaches into Neural Networks, Control Systems (And theory), Cybernetics, and probably a few other fields I can't think of at the moment. also Swarm Intelligence is pretty much the same as AI but (most of the time) in a different 'format', so I am under the assumption this proposal would also cover research relating to Swarm Intelligence.

That is why this proposal is different.

2

Cross validated is about data analysis. While artificial intelligence can include a good deal of this, it is a different field. Machine learning is different for two main reasons:

  1. As many other people have already said, machine learning and artificial intelligence is much broader. A neural network wouldn't be exactly on topic in cross validated, however it would be the purpose for this site. Also, while machine learning includes a great deal of data analysis, artificial intelligence also pertains to natural language processing, robotics, understanding images, and anything else that are features of an intelligent being, while data analysis

  2. Artificial Intelligence often has a different objective than data analysis. Creating code that gets better at something is different than interpreting information, even if sometimes, in practice, similar tasks are being performed.

You must log in to answer this question.

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