|
These are the fields I may be able to choose from for a masters degree (may because I'm not accepted yet  ) I want to get your opinions about which may be more interesting for me (I don't have enough insight to know for myself  ) by giving a bit info about me.
1. computer vision
2. image processing
3. information security & cryptography
4. embedded systems
5. robotics
6. data mining
7. machine learning
8. bioinformatics
9. semantic web
10. parallel computing
First of all I'm kind of a guy who strongly dislikes all-theoretical kind of studying. Like you read read read and enter an exam and solve the questions and pass the course, or read read read and then write a report-paper-homework. I'd like to have an implementation side of things I'm studying, where I can see that things "working". Otherwise it just feels like I'm carrying knowledge from course notes to exams (and forget most of them after couple of months) just to finish the school and get a degree. I had to do this kind of thing too much when I was studying for bachelor's.
Let me give an example. "Analysis of Algorithms" was one of my most hated courses. Because there was like 40 algorithms to learn (memorize), and then the exam comes with first question:
- write X and Y algorithm in pseudo-code, and draw them step by step for input A. (25 points)
It's not hard to do. You just have to memorize it. I hated this kind of things. Everyone is perfectly capable of finding their pseudo code in a book or in Wikipedia. Plus, even the best students forget it after couple of weeks from the exam. This is just "carrying" knowledge.
This is an oversimplified example but most of the courses are like this. You just carry theoretical knowledge in your head to get a degree. I would like to avoid this as much as possible.
Now I want to give some of my thoughts(feelings) about each area in the list above. Keep in mind that I didn't have a particular interest in any of them before, it may be because I don't have enough insight to get interested in one, or I may not be an academic person altogether . If I decide the latter, I won't tire myself to get a degree.
1. computer vision, 2. image processing : these two are closely related afaik, but they are listed separately in my university's webpage. They are a bunch of algorithms to apply to an image and try to recognize patterns, objects, etc. There was a graduate project about reading Sumerian tablets, which I found pointless because why you would read such an important archaeological find using a computer and not the experts?
3. This is a horse beaten to the death and more imo. I was never interested in how "hacking" works, unlike most of the CS students. From the undergrad courses I took it seems security literature is more or less established but security holes come from implementation errors or people making mistakes, not from the lack of theoretical knowledge. (boring)
4. Should be more about electronics engineering than computer, because in an undergrad course about this area, students were %90 ee students, and ce students failed miserably (I was among them too). I don't think I am capable of mastering this one.
5. I honestly don't have an opinion on this one.
6. Heavy-duty theoretical field, given by the most heavy-duty professors.
7. "Machine learning" name sounds extremely interesting but it involves mostly algorithms, which I have trouble memorizing, or learning if you prefer, but at the end of the day I will be required to write them down in exams.
8. I don't have knowledge about Biology apart from high school curriculum.I don't know how it may affect studying in this field though.
9. The professor in this field at my school is, let's say, a problematic person. Otherwise I don't have much opinion.
10. This feels interesting and I'm leaning towards it. It may have a dominant implementation aspect with a strong theoretical background.
Anyways, feel free to shoot even if you didn't read the whole thing!
|
I'm a huge fanboy of machine learning (and neural nets in particular). The basic theory is really easy, which makes it even more exciting, because neural nets are simple and yet very powerful tools. I could teach you how to write a neural net that can read handwritten digits in 30 minutes.
But I don't wanna talk you into something. It's super important to go for your own interests.
You could check out www.neuralnetworksanddeeplearning.com to get an overview.
|
How many of them are you supposed to choose? Computer Vision / Embedded Systems / Robotics / Machine Learning would be a strong combo if you're interested in things like autonomous cars or robots.
|
On April 28 2016 04:56 Scorch wrote:
How many of them are you supposed to choose? Computer Vision / Embedded Systems / Robotics / Machine Learning would be a strong combo if you're interested in things like autonomous cars or robots.
I can take one course from each one if I want to but the better idea is to specialize in one or two I think.
On April 28 2016 04:48 beg wrote:I'm a huge fanboy of machine learning (and neural nets in particular). The basic theory is really easy, which makes it even more exciting, because neural nets are simple and yet very powerful tools. I could teach you how to write a neural net that can read handwritten digits in 30 minutes. But I don't wanna talk you into something. It's super important to go for your own interests. You could check out www.neuralnetworksanddeeplearning.com to get an overview.
Wow! I shall read about it more. Like I said I don't have much knowledge about them to have a strong choice.
|
I'm no expert, and I only just started my career in IT (working for ~3 years now) but I think 3. and 6. are the most lucrative if you ever want to move into industry.
Of course, for studying (and overall life happiness) its important to have to right motivations which primarily should be where your interests lie and not for other external factors.
|
Data mining has a lot of real-world applications, especially in the era of social networking. It's mostly practical work, not theoretical. That'd be where I'd go given what you've stated here.
|
1. computer vision 2. image processing
Opinion based I guess, this leads to a particular part of the field
3. information security & cryptography
Is increasingly becoming more important. The information security part probably has to do with tons of documentation, not to mention that if you become a penetration tester, about 70% of the work is documentation and boring. Cryptography is wildly interesting and very profitable to know, however it's as you say "learning algorithms".
4. embedded systems 5. robotics
I think both of these are more engineering, robotics is quite interesting, but it takes alot of work to accomplish a relatively small thing. If you like to create an application from scratch and then mold it onto one of you homemade machines, it might be something for u.
6. data mining
I'm no fan of big data, it's probably the most boring part in IT, analyzing, learning how to "google efficiently", etc. In our school we used to call it "computing with ladies" (no offense meant), since somehow the design & data mining studies had much more women than the others... I don't like the fact that data mining has little to do with the understanding of computing/computers.
7. machine learning
Awesome subject, but very hard, I think you need a passion for this (or create a passion for it), because otherwise you will likely get insane from all the unlogical errors and ways machines might behave. As Beg said, things like neural networks are wildly interesting.
8. bioinformatics
Not my cup of tea, a friend of mine does this and he had to give up 3 years of his life to even get through his study. Very interesting, very difficult, assume that noone outside your field of study will understand a word of what you're saying.
9. semantic web
I think this market is pretty much saturated, if I had the option I wouldn't choose this, just because its too wide and shallow. Next to that, it is mainly "front-end" computing, where I'm a firm advocate of "back-end" computing, since the outflow of this knowledge is bigger than the inflow currently.
10. parallel computing
I don't know much about this subject, I think it is very interesting, but mathematically very hard. Currently it's above my comprehension.
All in all, the most important thing is that you love what you're doing. There is no good or wrong here, as long as you are not "sighing" at everything you have to do it should be good. Remember that you may want to work into that specific field the rest of your life, as long as you enjoy it, its good.
I'd rather get 2000$ a month and have a great time at my job and work environment,
than I would get 8000$ a month and have a shitty job,which I hate doing and my work environment is shit.
People who would rather get the 8000$ a month are either kids who don't work yet or psychopaths who'm are able to shut out emotions
|
France9034 Posts
First of all I'm kind of a guy who strongly dislikes all-theoretical kind of studying.
Uh, well there's gonna be some in every field at least. You'll have to eat up the maths and logic before being able to implement things (and understand them).
The two fields that you might like more with regard to that I think are parallel computing, and Image processing, which could be great combined with Computer vision. Also, if you like seeing things work, these are pretty nice to consider vs., say, security for example.
Another combination of related things could be bioinformatics, data mining and machine learning. These can have strong interactions as well, if you feel like combing through data (but involves algorithm, sometimes a lot).
|
|
|
|
|
|
EPT
