10 Answers
History.
With
all the hype for deep learning nowadays people are missing so many
great ideas that were proposed and even developed to a significant
degree in the past. The results that neural networks show on highly
specialized tasks like playing Go or Atari games are incredibly
misleading. They make people think that by some additional simple trick
the same algorithms can also do everything else.
They can’t.
There
were many failed attempts to achieve machine reasoning in the past,
performed by very famous people like McCarthy, Vinograd and Minsky.
Learning about their experiments and the conclusions drawn from them is
the main prerequisite for understanding what really needs to be improved
and where the possible solution can be.
All
that doesn’t mean skipping the math, but it’s not really the main thing
there. You will find out what you lack when you start reading von
Neumann and Turing and Rosenblatt and Shannon. Those guys weren’t
blinded by 100 million acquisitions in the news and their scientific
work was far more related to the actual problem than anything that is
being praised in the media. It’s worth noting that their papers are
supplemented with a decent amount of philosophical discussion, they try
to give clear and unambiguous definitions - it makes their work much
easier to understand.
Many paths were tested in
that era and many of them are forgotten now. Being well informed about
the foundation of the AI field makes one much more desirable
collaborator than, say, yet another DL enthusiast.
Also,
avoid the software. When you make yourself dependent on the particular
tool from the beginning, it shapes your worldview and you tend to miss
things you would otherwise notice. Conducting experiments is very good,
but knowing what and why you are doing that is more important.
Papers
from the 50s are a good start, but I’d not limit myself and basically
sift through all highly-cited papers by CS pioneers.
Basic
math and coding skills. But it really depends on what kind of
algorithms you plan to work with. I would say first select the
algorithms you're interested in, then explore the corresponding math
background you feel you need to understand them. Discrete mathematics
is the most widely used in computer science and artificial
intelligence, but is probably too broad for your needs. If you're
undecided on what branch of AI you want to study, the book Artificial Intelligence: A Modern Approach gives a very nice overview of AI and will help you spot the skills you need.
AI
is not like physics or maths it is too broad and largely unsolved.
There are no true experts in this area and things are currently changing
very fast in this field. What you need is just an overview of the whole
field then specialize, like jump into machine learning (ML) later and
develop from there, yes you can take a quick look at the history of AI for starters, you also need to know about state spaces and maths such as linear algebra and of course coding skills as well as differentiable calculus. You also need to know about set theories from maths of course as well as logical operations like in digital electronics. You also need to know induction and as well as probability theories.
My
favorite approach is to learn as you go, AI is too large to be
considered a specialized field, thus only when a particular algorithm or
algorithms seems important to the problem you are trying to solve
should you go deeper otherwise you might be too theoretical, this brings
me to another point.
Do you want to be an AI
theorist or a practitioner? These are two very different things, for a
theorist you need to know a lot more than is necessary about the field
itself thus you don't need coding or experimentations just theories upon
theories, this is best done by taking online courses or at Universities
for example.
A practitioner on the other hand
needs AI techniques to solve complex problems thus here you start with a
problem statement and view AI as just another set of tools or methods
to help you realize that goal. Thus for self-taught people I recommend
being a practitioner and it means getting down and dirty coding, you
need to experiment a lot and learn as you go, this way you will be able
to apply AI in a manner like no other. You will develop a unique
perspective and approach just like athletes have unique abilities
because of practice.
Then don't forget one very important prerequisite, passion,
without it you are probably wasting your time. It is important to love
whatever it is you are trying to achieve because everything is easier
that way.
Hope this helps.
At the end it is all about curiosity
and about how eager you are to learn about AI. Do you just want to
scratch the surface or do you want to dig deep into the subject. Also
from which perspective do you want to look at Artificial Intelligence?
Learning
about neuroscience might be helpful if you really want to learn and
imitate the brain and are interested in copying nature directly to built
AI. Here you might want to built your foundations with some basic
biology(what are cells?, how do they communicate?, metabolism?) and
elementary physics (mechanics and electronics) and later on combine
those two and move towards biophysics before transferring your
understanding of how the brain works into the computer.
Computer
scientist and mathematicians look at AI from a mathematical point of
view which can be a bit abstract from time to time. At the end it comes
down to find things to calculate and find fancy ways of making
calculations quicker. So it is more about about creating algorithms,
which might be inspired by nature but are at the end of the day just
models. So here you have to have strong mathematical and coding skills
as Faisal Hassan and Franck Dernoncourt mentioned in this post.
A
philosopher might ask different questions about AI. A philosopher might
think about ethics of AI, it’s implications on society and what it
means for humanity in general. You have to be able to really think
something through in detail and also doubting your assumptions from time
to time.
Originally Answered: What are the prerequisites to study artificial Intelligence?
- Motivation to study
- self-discipline, to actually study
- Patience, it’s a long way to go
Real answer:
I guess, if you ask that you actually wonder “can i do it?”. Everybody will just answer you that you can only know by actually do it. You can also wonder “How to do it?” or more specifically “Where to start?”, then for both:
I guess, if you ask that you actually wonder “can i do it?”. Everybody will just answer you that you can only know by actually do it. You can also wonder “How to do it?” or more specifically “Where to start?”, then for both:
- Register either Computer Science / Maths / Applied Math degree.
- Register Coursera’s more than famous Machine Learning course by Andrew Ng.
Then you can follow ML masters, or continue the web based learning and competitions (kaagle &co).
As
a general advice (started ML with quite a poor math background): don’t
be afraid of math. You will need time (a lot) to understand some
concepts but you may find the right friend/prof/online resources to
tackle it!
Just, DO IT :)
Originally Answered: What are the minimum prerequisites before learning artificial Intelligence?
Normally,
you don’t start learning any areas of artificial intelligence until
your Master Degree level and Doctoral Degree levels. You also typically
need either a Bachelor’s Degree in Computer Systems Engineering
(intensive hardware and software based training), or, a conventional
Bachelor’s Degree in Computer Science. These are the standard
prerequisites for what you are looking to do. Robotronics LLC | Facebook
Hii Sayam Sawai Thanks for A2A.
If I start prerequisites of Artificial Intelligence, then first will be the Interest in A.I.
A.I
needs lots of Research before developing any kind of Project.
Understanding the flow of data and Most Important the knowledge of
Algorithms. The techniques of classifying data, removing noisy /
unwanted data and every single thing about Data.
Keep in mind that A.I is based on 2 things. 1. Algorithms 2. Data
here,
Programming language doesn’t matter either you can implement in Java,
Python, Matlab or R. The key things are Data and Algorithms.
So Before starting with A.I make sure you have good understanding of data and Algorithms.!!!
E-wish you all the best..!
NC - The Fifth Horse
Originally Answered: What are the minimum prerequisites before learning artificial Intelligence?
Nothing,
really, just like with all questions like “what do I need to do before I
do X?”. Preparation is a form of procrastination. Just google
“artificial intelligence” (there LMGTFY), land on the Wikipedia definition of it (like this Artificial intelligence - Wikipedia),
then start reading. Whatever you don’t understand, just click on it and
find out what’s it about. Wash your eyeballs in random articles about
your desired (as well as adjacent) topic, day after day, until you find
out what you need to study.
Whenever you
contemplate a new subject, you are not aware of what you don’t know,
hence this question. You need to read and work a bit to find out. Then,
slowly and steadily, you’ll start building that foundation of knowledge
and, before you realise it, you’ll be able to answer a question or two
on the subject. At some point (some people say 10,000 hours but hey,
that’s just a quantitative expression), you’ll be able to master, or at
least have some relevant opinions on your topic of choice: Four stages of competence - Wikipedia
Good luck!
Originally Answered: What are the prerequisites to study artificial Intelligence?
Almost
all my education (or rather, almost all my valuable education),
especially on this topic has been autodidactic, so I may not be the best
person to answer this question, but I'll give it a try.
Firstly you need determination to really understand it. Lots of people I know understand what tool or algorithm does what, but don't understand, or at least can't explain why. This
makes them valuable from an enterprising stand point, they are
efficient when doing things that have been done before, but not valuable
from a development stand point; they can't develop new algorithms or
trying to solve problems that haven't been solved. So you must decide
what you want; research and development in new ai techniques or to be a
data scientist.
Secondly I have encountered
statistics and computer science concepts again and again. You could
study those topics to get a leg up. I might even suggest education
yourself on neuroscience, that's where I started and I feel it really
helped; after all it is the brain we are trying to emulate.
IMHO:
essential ones are - Computer Programming, algorithms and basic Maths.
good to have- Optimization and Linear Algebra
essential ones are - Computer Programming, algorithms and basic Maths.
good to have- Optimization and Linear Algebra
No comments:
Post a Comment