These are just a a few notes for new CS graduate students who are
working with me on their MS thesis research, or would like to do so.


Your most important first step will be to take CS 5761, Introduction
to Natural Language Processing, in the Fall semester of your first
year. This class is typically offered each Fall, and it is important
that you take it in your first semster. This will help you get
oriented to NLP, and will in the end save you a lot of time as you get
up to speed preparing to do your thesis research. This class will
count towards your CS electives in your major, and does  NOT duplicate
CS 8761, Natural Language Processing, which is a graduate level class
you will take (in the Fall of your second year).


Any CS graduate student working with me as their advisor will be
required to complete a Plan A Thesis. While there is a Plan B project
option, I have only allowed that in a few (two!) very special cases
where a student joins me only during their second and final year in
the CS program.


We will work together to identify a suitable research topic, but in
general it will be in the area of Natural Language Processing and/or
Computational Linguistics. You can get a very good idea of the kind of
research we do by looking at previous MS student's work :


http://www.d.umn.edu/~tpederse/masters.html


You should also look at the various software tools we have created
over the years, and even give them a try. Some of them have web
interfaces, others of them must be installed on a Linux system and run
from the command line. You can find those by browsing around my web
pages ( http://www.d.umn.edu/~tpederse ) If we have done our job well,
you should be able to install and use these packages simply by
consulting the documentation. Let us know how well we have met that
goal, and suggestions for improvement are always welcome.


Most of the software packages have mailing lists associated with them,
and you are encouraged to join any and all that you find interesting.
In addition, there is a NLP @ UMD mailing list that I ask all students
who work with me to join, and anyone else who is interested in what we
do here is surely welcome. This is a place where we make announcements
and point out issues in the news that might be of interest. Please
sign up here :


http://groups.yahoo.com/group/nlpatumd/


In addition to completing an original piece of research (ie your
thesis), you will be required to take four 8000 level CS classes,
plus two additional classes that may be either out of department or
within the CS department. One of your additional classes should be CS
5761, Introduction to Natural Language Processing, and the other may
be of your own choosing.


It is possible to take free electives while you are a graduate student
- this is included in your TA or RA funding in fact! These are
typically 5000 level classes (or below) that do not count towards your
degree (but you can still take them for credit and grades). This could
include areas of personal interest (such as Theater or Physical
Education), or other academic areas (such as Mathematics, Economics,
Linguistics, Psychology, or Engineering Management). As long as you
are doing well in your required classes and your thesis research, I
will very likely approve.


It will take you four semesters of hard work to complete your thesis.
We will start working in the first weeks of your first year, and we'll
continue through to the end of your second year. Your course load at
UMD is relatively light (only 6 courses over 4 semesters) - that is to
allow you to spend significant time on your thesis work each semester.


You should plan on finishing your thesis by May of your second year.
This will require you to work steadily on your thesis through your
four semesters at UMD. If you do this, you will certainly be able to
finish your thesis by May of your second year. This will enable you to
move on to whatever comes next most conveniently and without
distractions or interruption. Students who delay finishing their
thesis until the summer of their second year or later often have
unexpected difficulties in making the transition to whatever is coming
next. We will create a plan during your first year (in your thesis
proposal) that will give you a realistic schedule that allows you to
finish by May of your second year. It is really your responsibility to
make sure you both understand and follow that plan. Also note that in
some summers I may not be available to supervise or review your thesis
work, so if you do not finish by May I may not be available to you
until the following September.


Your thesis work will inevitably include programming (typically in
Perl or Python). Any code that is used to produce results that are
either published or that appear in your thesis *must* be released as
open-source. The motivations behind this policy are described in a
short piece that appeared in Computational Linguistics in 2008 :


http://www.d.umn.edu/~tpederse/Pubs/pedersen-last-word-2008.pdf


This philosophy is central to much of what we do, so please make sure
you over the above very carefully.


There are two required pieces of writing that you must complete as a
part of your thesis. The first is your thesis proposal (mentioned
above), which is normally completed by the end of your first semester.
This is a 3-5 page document that describes the problem you will be
working on, what approach you will take, how others have tried to
solve this problem, and what the overall schedule for the work is. The
second is the thesis itself, which should be completed by the end of
your fourth semester (May of your second year.) Note that we will
start writing your thesis well in advance of this deadline, doing
preliminary writing in your first year and then working on the thesis
throughout your second year.


Both of these documents must be original, and you will do all of the
writing yourself (with my guidance). It's very important that the
writing be done in a proper academic or scholarly fashion, which means
among other things that there must be no plagiarism in the work.
Please make sure you read and understand the following, and always ask
questions if you are ever in doubt about what is appropriate in your
writing:


http://www.d.umn.edu/~tpederse/Docs/A-Plagiarism-Case-Study.pdf


Also, please read the following, which includes some general thoughts
and tips on writing about research :


http://www.d.umn.edu/~tpederse/Docs/The-Art-of-WAR.pdf


The summer of your first year (between years one and two) is an
excellent time to pursue an internship or similar opportunity. If you
are an international student, it is important to realize you will need
to obtain authorization from the US government in order to work as an
intern (or participate in an off campus research experience such as a
summer school). This authorization is obtained via something called
Curricular Practical Training (CPT). You will need to get my approval
for the internship, and also register for a 1 hour Internship class
for the following fall. Additional info about CPT can be found here :


http://www.isss.umn.edu/fstudent/cpt.html


Also, if you are an international student it is important to
understand the regulations surrounding Optional Practical Training
(OPT). This is an authorization that allows you to work for 12 + 17
months after graduation. You can find more details about OPT here :


http://www.isss.umn.edu/fstudent/opt.html


Finally, please review the CS department graduate program brochure in
detail. All degree requirements are spelled out there, as are
expectations regarding academic performance:


http://www.d.umn.edu/cs/degrees/grad/brochure.pdf


Finally, I always enjoy working with graduate students, and in general
have found that students enjoy their work with me. I realize the above
may sound a bit formal, but this is just to make sure you understand
as much as possible about my expectations and how I manage things
before you get started. A good understanding of that will help you be
successful and finish a thesis you will be very proud of in a timely
and efficient way.


Please let me know if you have any questions about any of the above,
or anything else related to working with me. I usually respond to
email pretty quickly,  so please feel free to get in touch
(tpede...@d.umn.edu)


Last revised June 14, 2015

Reply via email to