Is Artificial Intelligence a subfield of Computer Science?

  • Thread starter Gjmdp
  • Start date
119
4
At first glance it does seem obvious that AI is a subfield of Computer Science for that programming is essential in the building of intelligent systems. However following that reasoning, Physics could be classified as a branch of Mathematics since Calculus is essential for its understanding. AI research does have a different approach to that of Computer Science, as it needs of empirical research, since no mathematical structure known is capable of reasoning. Computer Science is rather a formal science, as it studies algorithms and other related topics. It seems that it's somewhat disputed whether AI is a field on its own or not. I'd want to know if I am making some wrong assumptions. Is AI actually a subfield of Computer Science?
 
Last edited:
88
33
The original AI research brought together many disciplines because the approach was generally holistic in nature and focused on replicating the knowledge and reasoning aspect of how it was (is?) assumed our brains work. However, the recent focus on machine learning digesting volumes of data has tended to bias the field to mathematicians who can program in languages such as R and Python, or configure systems such as Storm or TensorFlow. Because of the programming aspect inherent in this, it is commonly viewed as 'computer science', but as you note, that is disputable and given that statistics permeates so much of the modes of ML, I feel it is beyond pure programmatic, "code this to get that" approach of what is typically taught as computer science.

Just for fun, I've attached some of the ML options and it's pretty apparent that this is not computer science in any traditional sense.

With that preamble aside, I'd ask the 'why' of your question. Understanding your context can help provide a more directed answer.
 

Attachments

119
4
The original AI research brought together many disciplines because the approach was generally holistic in nature and focused on replicating the knowledge and reasoning aspect of how it was (is?) assumed our brains work. However, the recent focus on machine learning digesting volumes of data has tended to bias the field to mathematicians who can program in languages such as R and Python, or configure systems such as Storm or TensorFlow. Because of the programming aspect inherent in this, it is commonly viewed as 'computer science', but as you note, that is disputable and given that statistics permeates so much of the modes of ML, I feel it is beyond pure programmatic, "code this to get that" approach of what is typically taught as computer science.

With that preamble aside, I'd ask the 'why' of your question. Understanding your context can help provide a more directed answer.
Well I was a bit confused about what is Computer Science and what is not. I have recently started studying "6.080 / 6.089 Great Ideas in Theoretical Computer Science" (Spring 2008 MIT OCW) by Scott Aaronson. In the first set of lecture notes, it does seem that he feels Computer Science studies any conceivable system (the brain, the universe, ....).
In lecture one Aaronson stated:
"
Computer science is not glorified programming. Edsger Dijkstra, Turing Award winner and ex-
tremely opinionated man, famously said that computer science has as much to do with computers
as astronomy has to do with telescopes. We claim that computer science is a mathematical set of
tools, or body of ideas, for understanding just about any system—brain, universe, living organism,
or, yes, computer. Scott got into computer science as a kid because he wanted to understand
video games. It was clear to him that if you could really understand video games then you could
understand the entire universe. After all, what is the universe if not a video game with really, really
realistic special effects?
OK, but isn’t physics the accepted academic path to understanding the universe? Well, physi-
cists have what you might call a top-down approach: you look for regularities and try to encapsulate
them as general laws, and explain those laws as deeper laws. The Large Hadron Collider is sched-
uled to start digging a little deeper in less than a year.
Computer science you can think of as working in the opposite direction. (Maybe we’ll eventually
meet the physicists half-way.) We start with the simplest possible systems, and sets of rules that
we haven’t necessarily confirmed by experiment, but which we just suppose are true, and then ask
what sort of complex systems we can and cannot build.
"

So it does seem in this sense that AI is quite a subfield of AI, however I haven't found anyone else this perspective about Computer Science. He is a lecturer at the MIT.

References: https://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-080-great-ideas-in-theoretical-computer-science-spring-2008/lecture-notes/lec1.pdf
 
10,334
3,866
The original AI research brought together many disciplines because the approach was generally holistic in nature and focused on replicating the knowledge and reasoning aspect of how it was (is?) assumed our brains work. However, the recent focus on machine learning digesting volumes of data has tended to bias the field to mathematicians who can program in languages such as R and Python, or configure systems such as Storm or TensorFlow. Because of the programming aspect inherent in this, it is commonly viewed as 'computer science', but as you note, that is disputable and given that statistics permeates so much of the modes of ML, I feel it is beyond pure programmatic, "code this to get that" approach of what is typically taught as computer science.

Just for fun, I've attached some of the ML options and it's pretty apparent that this is not computer science in any traditional sense.

With that preamble aside, I'd ask the 'why' of your question. Understanding your context can help provide a more directed answer.
Nice chart in the PDF where did you get it? It looks to be an iThoughts generated mindmap chart, is that correct?
 

anorlunda

Mentor
Insights Author
Gold Member
6,294
3,498
There are no language police in this world. If you are looking for logic and consistency in natural language, you will be disappointed. That applies to the natural language of science and engineering also
 
88
33
Nice chart in the PDF where did you get it? It looks to be an iThoughts generated mindmap chart, is that correct?
Think the chart was built in Freemind, gleaned from various sources, probably getting out of date by now, and certainly not the last word on the burgeoning options that make up machine learning :biggrin:
 
88
33
however I haven't found anyone else this perspective about Computer Science
Honestly, I think computer science is still a craft and nowhere near 'industrialized' in the way that maths, engineering and physics are. There are no "laws of computing" such that any practitioner can apply them and always get the same outcome as any other practitioner, and Aaronson is taking such an extremely broad church approach to his definition that it is hard to call him right or wrong on it.

But having worked with organizations large and small, I've found computer scientists to be considerably more pragmatic than Aaronson. They are just looking to get their hardware or software working because that's the gig and if that requires ML or NLP or an Expert System or just good old vanilla C# if else statements, that's what they are doing. They don't link their work to a top-down physics framework or talk of confirming things by experiment.

Indeed, to paraphrase one interpretation of Quantum Mechanics, they just "shut up and code" :wink:

Not sure if that helps or not, I guess I'm saying I've never come across a purists view of computer science of the kind Aaronson describes so perhaps I'm out of touch...or he's perhaps incredibly conceited!!
 
119
4
Aaronson is taking such an extremely broad church approach to his definition that it is hard to call him right or wrong on it.
Yeah that's what I thought. But since he's a lecture at the MIT, I thought that there would be more people that thought like him. But it seems there aren't. I guess AI is then a field on its own, as you guys have said, and doesn't quite meet Physics in any reasonable way, despite what Aaronson asserts.
 

atyy

Science Advisor
13,382
1,522
By convention, machine learning is a subfield of computer science.


By convention, parts of physics are also subfields of mathematics. The fields medals have been given for understanding physically motivated fomalisms like the KPZ equation and Landau damping. The Millenium Prize problems include questions about quantum fields theory and turbulence.
 
Last edited:

atyy

Science Advisor
13,382
1,522
So it does seem in this sense that AI is quite a subfield of AI, however I haven't found anyone else this perspective about Computer Science. He is a lecturer at the MIT.
He's now at Austin: https://www.scottaaronson.com/.
"I'm David J. Bruton Centennial Professor of Computer Science at The University of Texas at Austin, and director of its Quantum Information Center. Prior to coming here, I taught for nine years in Electrical Engineering and Computer Science at MIT. My research interests center around the capabilities and limits of quantum computers, and computational complexity theory more generally."
 
119
4
By convention, machine learning is a subfield of computer science.
It is, but Artificial Intelligence is much more than just Machine Learning, this is, Machine Learning is a subfield of both Computer Science and Artificial Intelligence.
 

Want to reply to this thread?

"Is Artificial Intelligence a subfield of Computer Science?" You must log in or register to reply here.

Physics Forums Values

We Value Quality
• Topics based on mainstream science
• Proper English grammar and spelling
We Value Civility
• Positive and compassionate attitudes
• Patience while debating
We Value Productivity
• Disciplined to remain on-topic
• Recognition of own weaknesses
• Solo and co-op problem solving
Top