6
When looking for specialized programming courses or with market demand, two of them stand out:
In my view, ML could be a subcategory of AI. So why is it a totally different course? It would have to do because they are very complex subjects or make more or less use of mathematical algorithms?
8
Attention: The approach of this post is just one of the ways to treat a totally broad subject like artificial intelligence. It should be understanding as objectively and briefly didactic and should not be interpreted as a closed entry, absolute truth or primary source. The Stackoverflow in English is a Q&A site objectives, and the elaboration of themes is not part of its scope . For more elucidative questions, treat in chat or subject-specific documentation!
Artificial intelligence is a multidisciplinary area involving psychology, philosophy, computing, mathematics, and as recalled by member Luiz Vieira’s contribution, also learned in other courses such as the economics course.
Machine learning is one of the subjects learned within the discipline artificial intelligence, that deals with approaches to computer learning algorithms. Learning is one of the known behaviors of the phenomenon of intelligence.
Follow the following reasoning::
- 1 - Men are made of atoms.
- 2 - Women are made of atoms.
- Conclusion - Men and women are of the same sex, because they are made of atoms.
This is a small demonstration of the fallacy of composition, which can induce wrong reasoning. This example is not to say that the subject of the topic is this, or that different approaches point to it, but a warning so that such related matters do not fall into fallacy. Therefore, from the semantic conceptual point of view, Artificial Intelligence and Machine Learning are different concepts. This does not mean that, depending on the approach, at some point, they can mean the same activity. It all depends on how it is contextualized. I hope I have contributed and not given the impression of being a closed subject.
Finally, I recommend the book "Artificial Intelligence" by Peter Norvig and Stuart Russell. In Brazil it is distributed by Campus. You can find it by Amazon.
You can also find this book in well known bookstores.
Hello Paul. Many of the AM techniques are learned also in other courses (mainly statistics and economics): linear regression, bayesian inference, etc. Therefore, although your answer is right, it is not necessarily canonical. I think it is worth editing to add this remark. Another thing is: the recommendation of the book is ok, but the final sentence seemed propaganda. I would remove this phrase to prevent your reply from being characterized as SPAM.
I edited out the final part. But wanted a suggestion of how to edit the initial part for the author of the question, and who has the same doubt, understand that these summaries are not absolute truths. You have editing privileges, could you improve that for me? If it’s not abuse on my part.
You can also edit friend.
I know, but I wanted a suggestion! .
Do you say edit in the AP question? I don’t think you need it. You can put any suggestion directly in your answer. I guess it’s part of the answer anyway.
6
An addendum, perhaps this image is more explanatory:
Image source: https://www.razorrobotics.com/artificial-intelligence
Machine Learning can rather be considered a sub-item of Artificial Intelligence, the two subjects are intrinsically linked, many concepts that learn in AI are then employed with ML. We can also make an analogy with Differential and Integral Calculus, or Discrete Mathematics for example that are sub-items of Mathematics, but there are many different concepts of what is learned in a pure Mathematics module.
The subjects are taught in different modules because there are many different applications and concepts.
I remember in college I learned this concept, which I believe to a certain extent:
Artificial Intelligence can be considered any non-human system that performs a certain autonomous task, without the need for intervention of a biological intelligence. Any algorithm that serves some purpose can be considered an artificial intelligence;
Machine Learning works as a system that learns which behavior to follow depending on the iterations it has. At first it has a default behavior, but that over time, depending on what you enter and the expected output, may or may not change this behavior.
The subjects are taught in different modules because there are many different applications and concepts. Thank you!
Excellent response. + 1
4
the view that ML is a subcategory of AI is correct, but to clarify this understanding, it is important that you know that when we talk about AI, we are talking about a huge variety of subjects, some of them are: expert systems, machine learning, search algorithms, etc.
In general, AI disciplines are concerned with presenting an overview of these subjects, promoting an understanding of what exists and the possible possibilities, and therefore the student/researcher/engineer/programmer will be able to choose which subcategory will investigate better. This can be done through more specific disciplines, such as machine learning, which seeks to bring a deeper knowledge of techniques, such as: Tree-based learning, Artificial neural network, Genetic algorithms, Clustering.
I don’t know who has negatively criticized your post, because it is correct. I found a good approach. It has my +1. It even complements my answer!
Opa @Paulosérgioduff, I also did not understand, but kkkkkk and you are right, yours was much more elaborate and in a way all who commented were complemented the response of each other. Thanks for the +1. D
I’m still going to comment on the Meta about Downvoters! The good news is that whoever has the same question and google it, will have rich and detailed information.
3
"So why is it a totally different course?"
This question is independent of how the machine actually works.
"It would have to do because they are very complex subjects or make more or less use of mathematical algorithms?"
There is no computation without algorithms that use mathematics... The amount or size is related to the algorithm itself, depends on what it will do, but both AI that will define a whole range of behaviors and interpretations as the ML that will absorb these definitions, will use a reasonable amount of complexity.
I have no doubt that these are closely related terms and that ML is a sub-category of AI.
The difference in my view, since ML is a sub-category of AI, is that AI is more linked to cognitive functions inherent to the human being, and that ML uses these concepts to learn and perform.
In my view it’s as if AI is the theory and ML is the practice.
So to answer your question, "Why Artificial Intelligence and Machine Learning are Different Subjects?" ...
They are not different matters, they are complementary matters...
Here you can see that ML is AI sub-category, of course it’s wiki, but logically we can agree that yes...
Machine learning or machine learning ("machine Learning") is a sub-field of computer science that evolved from the study of pattern recognition and the theory of computational learning in artificial intelligence.
And here can see a little more through the mind of one of the precursors of AI.
Great @Magichat, just note that I said "make more or less use" and not "not use". I said this because when seeing the requirements of these two courses, one asked for a level more advanced than another.
Ah yes. Rereading I could understand better, but the amount of algorithm is linked to the algorithm itself, both will use a considerable amount. From the answers about AI and ML indicated in the comment of your question, you can see that it is not short;
"I have no doubt that these are closely related terms and that ML is a sub-category of AI." I agree, and I even think most people agree. But there are those who disagree: "[...] Following this line of Reasoning, Machine Learning is NOT a subset of AI. It really is the ONLY Kind of AI there is."; and also "Machine Learning doesn’t care about Intelligent behaviour but Accuracy and Patterns.".
@Luizvieira I understand, to refute not only is valid as necessary, however I believe that using logic my statement becomes irrefutable, I do not believe there is ML without the understanding of AI, and I believe that AI is an earlier term...https://en.wikipedia.org/wiki/Journal_of_machine_learning_research , but I could be wrong...
Man, like I said, I don’t disagree. But it is that it depends a lot on the interpretation of the individual, especially in the light of the problem he is dealing with. That image (which is mentioned there in the link I mentioned) illustrates this his "irrefutable logic". But at the same time she illustrates how this is all so multidisciplinary that I wonder if arguing about this is not discussing the sex of angels. Perhaps it is more useful to discuss whether a particular technique belongs to area X or Y (and yet I have my doubts).
@Luizvieira did not say that you disagreed, only said that the link he sent in a way refuting the idea is that it is refutable (and was refuted with this image he refers to). I don’t think it depends on interpretation, and as I said I don’t see how to separate ML from AI, but I see AI without ML... hehe but I understand your concern about utility, but then the question would be another,,, Ussa seems that I learned this word hj (refute)...kkkk.. There are times that the formality, makes me seem silly... Good morning man....
Browser other questions tagged artificial-intelligence machine-learning
You are not signed in. Login or sign up in order to post.
A machine that learns would not be intelligent? And the intelligence of a machine, would not be artificial?
– Marcus Becker
I don’t have much knowledge on these aspects, but I see that AI is a very comprehensive discipline, that MA is within this discipline, MA is more centered on theories of how a machine learns and the standards it can recognize.
– Leonardo
Vixi, colleague... I think this question tends to be based on opinions, huh? In my opinion you are right, in the sense that matters are related (and that AM is, in a sense, a discipline of AI). But there may be those who disagree, perhaps because they see the applied knowledge of MS more as having origin in statistics, for example. The choice to have a separate course may stem merely from the amount of subject to be dealt with (since both areas are enormous). Anyway, I wonder how constructive this particular discussion will be. :/
– Luiz Vieira
@Luizvieira I agree that opinions should be avoided and respected, however in this case the evidence of the terms points to a specific direction...https://pt.wikipedia.org/wiki/Journal_of_Machine_Learning_Research I do not send this link because it is wiki, but by the publisher...
– MagicHat
@Luizvieira anxiously awaits one of his magnificent answers on the subject haha
– MarceloBoni
haha Thank you for the consideration, @Marcelobonifazio. But the answers have already been given (including in comments) by the other colleagues. And the question has just been closed (automatically, by the way). I like the subject, but I think the closure is just because this discussion is difficult (and quite opinionated, in my opinion rs). Why are Mathematics and Statistics different subjects? Why do you study Logic in Engineering, Computing and Philosophy? See how these questions are very similar. Anyway, even if it wasn’t opinionated, I hardly think it would be constructive for this site.
– Luiz Vieira
Curiously, today UOL published an article about scientists using data to predict everything. This type of process (in which data is used to build a predictive model) is certainly an important part of AM. Note that although the article does not cite AI once, it cites "mathematical models", "statistical model", and "[...] mathematics and statistics are the ingredients of the perfect cake recipe".
– Luiz Vieira
All the answers just made me think that a course sold only as AI is wrong. Please specify which topics of the MI this course addresses.
– Marcus Becker