Hello,
My situation is a little strange because I am a Statistics PhD student but I ended up doing research in Natural Language Processing (NLP)/ Deep Learning due to my supervisor's recommendation.
I feel strange because although I can read the NLP paper, understand them, and come up with my own research topics (so far I have identified 3 different research topics), I am not a part of computer science department NLP lab. My supervisor into applying/developing machine learning algorithms for analyzing open-ended questions, so I guess his work is somewhat related to NLP, but he is not really an expert in NLP/Deep Learning. I do like my current supervisor, he is trying to help me out in the best way he can, and I want to continue work with him. But I am wondering,
- Is it advisable for me to seek a professor from computer science NLP lab to be my co-supervisor (upon my supervisor's consent)?
- What are the advantages of doing a NLP research in a big computer science NLP lab?
- Is collaboration important for a PhD student to do research in NLP? I see many PhD students who publishes at top NLP conferences are often co-authored with multiple number of collarborators. I am just doing NLP research with my supervisor, and my supervisor and I don't really have any connection with NLP researchers. Should I make an effort to find NLP researchers whom I can work with for my research (of course, I will be doing most of the work since I want to be the first author)?
- If I need to seek collaborator / co-supervisor who are familiar with the field of NLP, how should I approach them? can I try sending them emails and see if they are interested in my research? I guess the best way is to talk to people at a conference but due to COVID-19, everything is taking place online and I doubt whether I will be able to make any connections.
- Do PhD student from a big NLP lab have better computational resource than the PhD students who does NLP research outside of those labs? My supervisor recently set up an account for the national supercomputer so that I may take an advantage of it, but since almost every researchers in my country have access to this resource, when I submit the job on SLURM the queue wait time can be long. Are PhD students from computer science NLP lab often free of this problem? (do they have a access to better computational resource?)
...Lots of questions! Could someone advise me on these issues?
Thank you so much for your time,
That is indeed a lot of questions Hyunjin Cho but it is normal when you start your Ph.D. I was filled with so much questions, when I started my own Ph.D.
- If you have background in Statistics that is really a plus point in NLP domain, IMO. it will be easier to understand the mathematical foundation behind the algorithms being used.
- It would be an advantage to get co-supervisors from NLP domain if possible. However, I would not suggest to have multiple co-supervisors as well :)
- To seek for collaborator, finding them online, reading their papers are really good way to do it.
- You need a OKAY computational resources. Most of the time, you really don't need very sophisticated computational resources. Of course, it is an advantage if you have it. I have a normal laptop with 32 GB Ram, and 2070 RTX GPU. It servers well for me.
I am very interested in computational linguistics. I am from Computer Science background and recently digging my opportunities in NLP. I hope, I was able to answer some of your question. Feel free to connect if you feel so.
Regards, and Good luck
If you have expertise in Mathematics/Statistics, you do not need to worry about. In such a situation you are the wright person for any type of AI/ML researches. You will be very Okay after a year & half.
In my country, the preparation of PhD thesis should be as part of Oficial research. In this regard, you need to take part in the official research of NLP laboratory in the NLP area. On the other hand, this participation possible for outsider specialists not only for NLP Laboratory members.
- Badges
- Science topic
- Similar topics
- Computer Science