SoP for applying CS PhD program with research experience that is not my interest

Question

I am applying for a CS PhD program in the US, next year, as an international student. I will complete my CS undergrad in a few months. I already have research experience with my project advisor. Also, I did a research internship at another well-known school with a similar topic. This project was presented in an international conference, too. I think this should be beneficial for my application.

However, I am not that interested in my research topic. It is related to computational neuroscience. At first, I just wanted to gain experience working with Machine Learning but I did not have an idea what problem I should be working on. Neuroscience sounded interesting to me, so I picked that sub-field. I did apply ML in my research project as I wanted but the neuroscience stuff was definitely not my thing. It has been an exhausting time for me to work on a project I do not like. I am still OK with ML and I am passionate about it. The thing is that I want to get away from neuroscience when I apply for a PhD. Right now, I have an idea to work with ML on low-power IoT device (such as camera) but I still have time to think about it thoroughly.

The question is that how can I write a strong SoP in this situation? If I write all about ML & neuroscience in my SoP, I am afraid that I will end up in a lab that I do not like. On the other hand, if I mention ML & IoT, I may have a weaker SoP since I have no experience in this sub-field. From my understanding, SoP is not a commitment but I am not sure that this will affect my journey or not.

Answer 1

In my own SoPs I wrote about my past experience and how it informed what I wanted to do in the future. For me in particular, I had done some computational research and some bench work unrelated to neuroscience, but I wanted to do some more wet biology in neuroscience. I wrote about how those relatively unrelated pieces of experience would apply to what I wanted to do next. I think it makes a lot more sense in a statement of purpose to talk about what you want to to study and learn rather than only what you've already done. If I said "I've been growing walnuts my whole life and I want to do this PhD to grow more walnuts" - that might be great in a job application to be a walnut-farmer, but if I read that in someone else's SoP I'd wonder why they were after a PhD and if they had any idea what a PhD actually was.

Presumably in your work with machine learning in computational neuroscience, you've learned about some of the benefits and limits of machine learning in that domain. You now know a bit what it's like to work with those tools, and you'd like to apply them to a new problem. Presumably, in a CS PhD, you'd be expected to do more work developing the tools rather than just applying them. That seems like a pretty straightforward SoP to me.

Answer 2

For CS doctorates in the US, you don't normally need to come with a research topic. This is especially true for someone coming straight from an undergrad institution. You can start fresh. (Note that "undergraduate" has different meanings in different places.)

However, if you apply to a program that, itself, has a strong neurological science focus then you will probably get stuck there. But there are a lot of possibilities outside that area. Most such US programs expect incoming students to have a broad education with some focus on a field, but not an especially narrow focus.

Make your SOP forward looking for the most part. The CV tells what you have done already, make the SOP about what you want to do: Machine Learning or whatever. Talk about why you can be expected to succeed in the new endeavors.

In some other places, students may need to already have a strong research focus to begin. That isn't true here. Expect that most PhD programs will include some coursework.