Research opportunities (in machine learning) for undergraduate students at colleges without good research programs

Question

I am fascinated by machine learning, and have been experimenting with ML techniques in code for a few years now. I would really like to run an experimental study about the most effective ML systems for a certain problem, and incrementally improve the model. Basically, I'm looking for research, but my college doesn't have well-defined research programs. Are there opportunities to do research with other institutions for undergraduate students?

Update:

I see it's now worth mentioning that for various reasons, I attend a community college. Hence, no local research opportunities.

Answer 1

One thing to consider is research experience for undergraduate (REU) placements over the summer. Many of these are actively built on affording research opportunities to students in schools that do not necessarily have their own research infrastructure.

Here is a list of those funded by the NSF, but this is by no means all of them: http://www.nsf.gov/crssprgm/reu/list_result.jsp?unitid=5049

Answer 2

i would suggest that faculty advisors and supervisors at other undergraduate institutions have their hands full with their own students. if you had a preexisting relationship with someone it might be a different story. however, you don't have to have a mentor to do research. expert help will make for better results and can help with regard to getting published. there is undoubtedly a large literature here on SE about cultivating mentorships.

here are several options: request some independent study at your school, which means you get an advisor. poke around online and in meetups to find project collaborators; enroll in an online data science program; do what you can on your own until you can apply to a graduate program in CS, information, or statistics, where you would have opportunities for at least some research advising. in all these cases, having an idea what topics you would like to work on will make the commencement of useful projects more likely. good luck.

Answer 3

Look at summer internships in companies like Google, Facebook, Microsoft, IBM... (they have their machine learning groups).

If your schools does not do good machine learning - it would be both hard to do it there and get admitted at research programs at other universities or institutes (if your advisor it's not know there, you are unlikely to get accepted).

Answer 4

Open-source communities can be very welcoming to contributors.

Testing and benchmarking a tool like sklearn is not easy, and they for sure will like someone to study the effectiveness of methods implemented and maybe provide some improvements (or at least the foundations for quickly testing any new algorithm or change).

There is a substantial amount of people there with academic background that can assist you with having a research direction. They won't be able to give you college credits, though (but you may be able to find someone in academia that could write a letter - maybe your college can give you credit if the project was supervised by a external academic)