1.
Introduction
Welcome to Area I!
This guide is primarily written for entering graduate students who
have come from undergraduate universities other than MIT, and who
intend eventually to obtain a Ph.D. at MIT. It is intended as a planning
guide for the first two years, which is the normal period for obtaining
the masters degree and completing the departmental Technical Qualifying
Exam (TQE).
For administrative convenience, the faculty and graduate students
in the department are divided up into six primary research “areas”
according to their preferences. And although the area definitions
are somewhat arbitrary — and the boundaries between them often
quite artificial — many of these areas have a long history and
well-established culture. Area I is certainly no exception, and as
a result its boundaries are deliberately fuzzy and there is considerable
overlap and many connections with other areas. Each area is responsible
for a portion of the graduate curriculum that reflects its intellectual
core. Over the course of their graduate careers, students take courses
both within their chosen area and in other areas and departments to
fulfill their degree requirements, and end up interacting with a similarly
broad collection of faculty, staff, and students.
Area I research concerns itself with a broad spectrum of problems
of communication and coding, systems theory and control, optimization,
statistical inference and decision theory, and signal processing,
as well as the shared methodological underpinnings of — and
increasingly the interactions between — these different fields.
Research topics range from fundamental principles to application,
from analysis to synthesis, and from theory to experiment and simulation.
Thesis research can involve different combinations of the above, depending
on the student's interests and the nature of the problem. A list of
several recent Area I graduate theses is provided at the end of this
document as examples.
In general, Area I draws graduate students from a broad range of backgrounds,
and with a wide variety of objectives for graduate study. It is the
culture of MIT to encourage students to take the initiative to tailor
their graduate program accordingly.
2. Initial Priorities for New Graduate
Students
Our graduate program involves a combination of research, course work,
and, in most cases, teaching. A number of requirements must eventually
be satisfied. Experience has shown that the first-year student will
do well to focus on the following primary goals:
·
Finding a masters thesis supervisor
· Ensuring fluency in the necessary undergraduate background
for Area I
· Successfully completing at least two Area I core graduate
courses with mastery
Completion
of the technical qualifying exam (TQE) will follow from success in
the latter two areas. Experience has also shown that the first-year
student should not focus too much on the following secondary goals:
·
Racking up credits (no one has ever failed to obtain an MIT degree
for lack of credits)
· Sampling a tempting smorgasbord of interesting courses
· Beginning your minor requirement
· Satisfying your TA requirement (unless you are supported
by a TAship)
3.
Finding a Thesis Supervisor
Finding a thesis advisor that is a good match in terms of interests,
style, and temperament should be the student's top priority from the
moment of arrival, if not earlier. As a result of the popularity of
Area I among students, there are many more graduate students in Area
I than can be supervised by the Area I Faculty alone. However, there
are many faculty and staff whose primary affiliation is an area other
than Area I, but who have Area I oriented research projects and supervise
Area I graduate students. So, while the opportunities are plentiful,
the process of identifying and picking one is less straightforward
than in other areas.
Students supported by a research assistantship (RA) have already achieved
this goal. Such students should register for 24 units of 6.991 (Research
in EECS) or 6.ThG (graduate thesis research), plus at most two 12-unit
subjects. (Note that in terms of strategy it is perfectly acceptable
to register for more subjects initially, and then to drop all but
one or two after sampling the first few weeks of lectures; in general,
it is very easy to add and drop subjects at MIT.)
Students supported by a teaching assistantship (TA) or a fellowship
(or other support) should start immediately to try to find a thesis
advisor. Such students should register for 6.961, a 12-unit subject
that introduces the student to graduate research in the department.
The first phase of this subject involves finding a research supervisor
for the term with an interesting introductory research project. Note
that while most 6.961 projects evolve into masters theses, the masters
thesis can ultimately go in a different direction and even involve
a different research supervisor, depending on opportunities. TA’s
also typically register for 24 units of 6.981 (Teaching in EECS),
leaving them room for one additional 12-unit subject. Students with
fellowship support usually have room for two such subjects.
Finally, some hints:
· If there is a particular professor with whom you have a
strong desire to work, a good strategy is to take a course they
teach and do very well in it.
· Be open to research opportunities in Area I that may be
less familiar. While incoming students see examples of Area I fields
and topics as undergraduates, many important and exciting themes
within the area are not encountered in the undergraduate curriculum.
Learning about those through the papers and web sites of area faculty,
groups, and labs, should be a priority for new graduate students.
· The graduate office has a number of resources for helping
students find suitable research supervisors. One of particular note
is the EECS brochure Research Interests of Faculty Members Who
Supervise Graduate Theses, which every new graduate student
should read through.
· As in many other dimensions of graduate school life at
MIT, our environment favors those with a strong sense of initiative
and perseverance, and creativity and resourcefulness play key roles
in finding a good opportunity. Area I graduate students frequently
find interesting research opportunities in any number of different
departments and laboratories on campus, as well as in laboratories
off-campus including Lincoln Laboratory, Draper Laboratory, and
a variety of medical and biomedical laboratories.
4.
Undergraduate Background
Most of the Area I graduate subjects have a strong mathematical bent
and require not just an exposure to, but a fluency with, undergraduate
background in linear systems, probability, and linear algebra. Students
should assure themselves that they have such fluency as a solid foundation
for their graduate work.
A common mistake is to enroll in an Area I graduate subject thinking
that you can correct your background deficiencies as you go along,
and then to find out about the time of the midterm exam that you are
falling behind and doing poorly because of your inadequate background.
It is much better to correct your background deficiencies first (and
you will be much happier).
Some students can achieve mastery of background subjects by self-study,
and some can’t. If you attempt this route, you should complete
your self-study before the term begins.
There is absolutely no dishonor in taking MIT undergraduate courses
to bolster your undergraduate background. This is very common in the
Area and highly recommended if your undergraduate institution didn’t
emphasize this background enough. Relevant courses at MIT include:
· 6.003 (Signals and Systems)
· 6.431 (Applied Probability)
· 18.06 (Linear Algebra)
· 6.011 (Introduction to Communication, Control and
Signal Processing)
5.
Core Area I Graduate Subjects
Area I offers a broad range of graduate subjects in its constituent
fields, most of which can be taken with few prerequisites other than
that the common undergraduate background discussed above.
Certain of these graduate subjects are regarded as “core,”
in the sense that they are appropriate for first-year graduate students
with adequate undergraduate background, and are prerequisites for
further subjects in the same area. These core subjects include:
· 6.241 (Dynamic Systems and Control)
· 6.251 (Introduction to Mathematical Programming) or
· 6.255J (Optimization Methods)
· 6.262 (Discrete Stochastic Processes)
· 6.341 (Discrete-Time Signal Processing)
· 6.437 (Inference and Information)
· 6.450 (Principals of Digital Communications I)
· 6.972 (Algorithms for Estimation and Inference)
Most first-year graduate students should plan to take at least two
of these core subjects in their first year.
All of these subjects assume a fairly high degree of mathematical
maturity and involve a relatively heavy time commitment. It is always
better to take fewer subjects each term and do well in each than to
take more subjects and do mediocre work in all of them.
6.
The Technical Qualifying Exam (TQE)
The TQE is a departmental exam which involves proving competence in
four subjects, as follows:
· Demonstrating mastery of two EECS undergraduate subjects
by taking exams, normally at the end of the spring of the first
year. Area I students most typically take the exams in 6.003 (Signals
and Systems) and 6.041 (Applied Probability), though this is not
a requirement. The graduate office maintains the list of choices.
· Passing two eligible graduate EECS subjects with high marks.
Note that most (but, perhaps paradoxically, not all) of the core
Area I graduate subjects listed in the previous section can be used
to satisfy the graduate portion of the TQE. When planning your first
choices, check with the graduate office to be sure which are eligible
for TQE designation. TQE Subject List.
The undergraduate exams are graded S (Satisfactory), M (Marginal),
and U (Unsatisfactory), which may be regarded as corresponding A,
B, and C , respectively. The TQE is passed immediately and no oral
exam component is required if a student gets either 4 A’s, or
3 A’s and a B. If a student does not meet this grade threshold,
they proceed to the next portion of the exam, in which follow-up oral
exam(s) are taken in the area(s) of identified weakness until adequate
competence is demonstrated. It should be emphasized that there is
no shame in taking an oral exam; until relatively recently in our
department’s history all graduate students took an oral exam.
The current TQE format simply removes this component of the exam for
the subset of students for whom there is already evidence that their
backgrounds are sufficiently sound to proceed.
7. Beyond the Core Subjects
The student may choose as many additional courses as he or she can
handle in order to obtain greater depth in the various topical subareas
of Area I, greater breadth in other areas, or just to satisfy intellectual
curiosity or be exposed to a great teacher. However, keep in mind
that:
· Again, it is to do well in fewer courses than not so
well in more courses
· There will be plenty of time at MIT to take all the
courses your heart desires
The MIT Bulletin, your graduate counselor, your research supervisor,
your instructors, and your graduate student peers are all excellent
sources of suggestions for subjects that may be a good match to Area
I students and interests. While Area I students invariably end up
taking at least as many subjects outside Area I’s offerings
as they do inside, in the remainder of this section we summarize some
of the historically popular Area I follow-on offerings:
· 6.231 (Dynamic Programming and Stochastic Control)
· 6.242 (Advanced Linear Control Systems)
· 6.243J (Dynamics of Nonlinear Systems)
· 6.245 (Multivariable Control Systems)
· 6.252J (Nonlinear Programming)
· 6.263J (Data-Communication Networks)
· 6.291 (Seminar in Systems, Communication, and Control
Research)
· 6.343 (Digital Speech Processing)
· 6.344 (Two-Dimensional Signal and Image Processing)
· 6.345 (Automatic Speech Recognition)
· 6.433 (Recursive Estimation)
· 6.435 (System Identification)
· 6.441 (Transmission of Information)
· 6.442 (Optical Networks)
· 6.451 (Principles of Digital Communication II)
· 6.452(Principles of Wireless Communication)- · 6.454 (Graduate Seminar in Area I)
· 6.455 (Sonar, Radar and Seismic Signal Processing)
· 6.456 (Array Processing)
8.
For More Help
In addition to discussions with their graduate counsellors, new students
are free to consult any other faculty members in Area I for guidance
in selecting appropriate subjects for their initial semesters, or
for assistance in dealing with any difficult or awkward situations
they may find themselves in academically in the course of their program.
More senior graduate students are also an invaluable source of assistance.
Finally, the most useful resource you are likely to find is the Department’s
Graduate Office itself, which has a wealth of information as well
as an extremely dedicated staff. In particular, Professor Arthur Smith,
Marilyn Pierce and Peggy Carney can be counted on for an attentive
hearing, wise counsel, and important practical help.
9.
Area I Faculty Members and Affiliated Faculty/Staff