2

I read somewhere online, how one shouldn't show their interest in genetic algorithms to a potential grad school, as its not the 'in' thing in todays CS community, compared to Machine Learning algorithms. I am interested in genetic algorithms and reading such a notion left me a little bothered, honestly.

closed as primarily opinion-based by scaaahu, Brian Borchers, David Richerby, JeffE, Florian D'Souza Dec 12 '17 at 16:03

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • keyword for you is Big Data analysis aka. bioinformatic, yes you right, Machine Learning is better for CS, and today market – SSimon Dec 12 '17 at 6:22
  • Is your interest focused on developing new genetic algorithms, or would applications of genetic algorithms also be of interest for you? As for applications, there is a vibrant community that applies optimization algorithms to software problems. Search-based software engineering would be the keyword here. – lighthouse keeper Dec 12 '17 at 7:29
  • I would like to keep away from really specific and focus interests. Primarily because I am just not sure so early being in undergraduate program. But yes I do like applying genetic algorithm to a particular problem. – nocturnal_study Dec 12 '17 at 7:37
  • 1
    @lighthouse keep I will remember to check that out. – nocturnal_study Dec 12 '17 at 7:54
  • 1
    When you say "genetic algorithms" do you mean: algorithms for genetics, or do you mean the nondifferentiable optimization technique? If you mean the second, then it really depends on the ML program. A statistical ML program will likely not respect that due to the nondifferentiable aspect. – Andnp Dec 12 '17 at 16:50
7

When applying to a school you should be aware of the faculty at that school and the research they perform. Take a look at their research papers. If nobody at the school does work in genetic algorithms, then it probably isn't going to help you to mention your interest in the area.

If you have a sincere interest in genetic algorithms, then you need to find a school with researchers that are doing work in that area. If there is a prominent researcher at a school working on genetic algorithms, then you really want to mention your interest. If the researcher is accepting new students it might improve your chances of acceptance.

There are always hot areas in a field, and machine learning is currently hot, and genetic algorithms aren't as hot. But, there is good work being done in both fields. (There is also bad work being done in both fields.) I would advise you to find someone doing excellent work, and then the field that you are in will not matter as much.

  • 2
    This is vastly more important than what is hot or trendy. If you aren't interested in what the faculty at the school are working on, it's not just a bad sign for you getting into the school, but also a sign you wouldn't be happy there. – AJK Dec 12 '17 at 7:06
  • 1
    Like I mentioned, I just am not too sure about my specific interests in research so early. My research has been in genetic algorithms, so I have to talk about that in my SOP as that is all I have as of now. My concern is primarily because of this...So to paraphrase I haven't taken up this particular topic just because I absolutely love that more than anything else, it just happened to be one of the initial choice. On a sidenote, I did enjoy working on it though. – nocturnal_study Dec 12 '17 at 7:40
  • @nocturnal_study - I think you are right to not limit what you are going to be interested in. If, when you do your application, you just explain why you have liked your genetic algorithm research so far, and make a convincing case that it taught you a lot which could be applied further, that sounds like a strong rationale to join a number of different sorts of research groups. – AJK Dec 13 '17 at 6:25
1

I suggest having a quick look into combining genetic algorithms and neural networks, some work has been done in this area, perhaps there is more to do. The trend seems to be combining different machine learning techniques (e.g., deep reinforcement learning) and this is something the original poster could exploit to have their cake (research GAs) and eat it (submit to machine-learning conferences where neural nets are hot).

The question distinguishes between genetic algorithms and Machine Learning. Yet, a quick search on genetic algorithms and machine learning shows that the former may be (and often is) considered a sub-discipline of the latter. Given that the two have (some) overlap there may be an alternative.

Not the answer you're looking for? Browse other questions tagged or ask your own question.