4 ms·

I plan to enter a PhD program in 1-2 years to specialize in ML/Deep Learning. Assuming it'll take 5-6 years to complete my degree how applicable should my skill

• adrenalinelol 10y ago
  No one can tell you for sure, if it's your passion do it, if you're hoping for a big payday, I'd reconsider.
  • whorleater 10y ago
    I don't actually think that's true, if the "AI bubble" bursts at some point in the near future, the people who'll be in trouble with be those without formal education to back them up.
    • toomuchtodo 10y ago
      > the people who'll be in trouble with be those without formal education to back them up. The people who can't hack it are those who'll be in trouble. Tech has never much been the place where credentials are necessary. Don't specialize and saddle yourself with years of college debt if you're unsure of the field's long term prospects.
      • nwjtkjn 10y ago
        FWIW, one does not accrue debt doing a PhD in something like Machine Learning.
        • toomuchtodo 10y ago
          Depending on the program, possibly. But a PhD is not required to learn nor work in machine learning.
        • randcraw 10y ago
          The debt incurred in a wasted PhD is not monetary; it's lost time.
      • robotresearcher 10y ago
        Almost every good place pays their CS PhD students enough to squeak by without getting into debt. I wish this was more widely known. And companies like Google are pretty keen on academic credentials. They've assembled what must be one of the largest collections of PhDs in history.
        • komaromy 10y ago
          > Almost every good place pays their CS PhD students enough to squeak by without getting into debt. I wish this was more widely known. Right. The debt problems that people have after PhDs are more often due to their undergrad loans sitting around for 4-7 years while they were earning enough to subsist and not more.
      • rampage101 10y ago
        Depends where you live. In the USA credentials matter a lot... in Europe or else where they could care less if you studied somewhere.
    • nilved 10y ago
      That's never really been how the industry works. Experience is more valuable than education, so you should get experience while you still can.
• breakds 10y ago
  If you are aiming at the industry, I would say getting a research-oriented master's degree is much more cost-effective. Besides you can always learn the skills you need in your work.
• wolfram74 10y ago
  A PhD is, by it's nature, rather self structured. How much value will be added to your skills over that time will depend in no small part to how you spend it.
• godmodus 10y ago
  you'll be a programmer - that is what counts. How good of a programmer you will be will determine your success. never put your eggs in one basket (not saying you shouldn't become an ML expert though, that's pretty damn nice). as a Phd, you are probably good enough. as to ML, its adoption is hyped. it is powerful, but not as anyone really talks about. support vector machines and Bayesian learning have been around since the 70s/80s (ninja edit: SVM's since 1963! Markov Chains 1950s, Bayesian Learning/Pattern recognition sine the 1950's), but adoption has been slow due to the nature of business, which is now drooling over it since neural networks beat a few algorithms. due to the hype, more business will opt for ML now, but the craze will plateau and ML will become another tool in your arsenal. so basically, you really have nothing to worry about - use your Phd to do interesting things, come up with novel and new research and/or develop your own product. don't let your job security worries get in the way of enjoying what you want to do now, you're already good and in STEM (and if you don't feel good enough, work on yourself until you do).
  • cerrelio 10y ago
    > support vector machines and Bayesian learning have been around since the 70s/80s (ninja edit: SVM's since 1963! Markov Chains 1950s, Bayesian Learning/Pattern recognition sine the 1950's), but adoption has been slow due to the nature of business, which is now drooling over it since neural networks beat a few algorithms. This is one of the things I find hardest about convincing managers and leads of. They think things like CRFs and Markov models are "new" methods and too risky. So they opt for explicit rule-based systems that use old search methods (e.g. A*, grid search), which hog tons of memory and processor. Those methods rarely ever work on interesting problems of the modern day. They can understand the rule-based methods easily. They have a hard time leaping to "the problem is just a set of equations mapping inputs to outputs, and the mapping is found by an optimization method."
    • godmodus 10y ago
      I explain it using the infinitesimal method, which if done right using the hill climbing metaphor, often delivers. But it does take away the magic of "wooo, neural" :p
• cerrelio 10y ago
  The important thing about the PhD is that you've become an expert in conducting experiments and research. It really doesn't matter what ML techniques you've done, as long as you know everything else associated with building those types of systems. Just hone your research and experimentation skills and you'll be fine. I do have one suggestion: learn to handle dirty data. I work with ML researchers and notice two things: they're pretty bad software engineers (no knowledge of software patterns, bugs galore), and they almost never know how to clean their data. The latter is because they do a lot of their research using pre-cleaned, standard data sets. You never get that in industry.
• UncleMeat 10y ago
  If your plan ahead of time is to go into industry then a PhD is not a really great plan unless you will get a lot of personal satisfaction out of research. PhDs are incredibly inefficient from a professional development for industry perspective.