The PhD Programme in the Computer Science Department
- Goal Statement
- The Programme
- Method of Assessment
- Transfer from MPhil to PhD
- The Structure of a Thesis
- Getting Ready to and Actually Submitting
- The Role of Key Staff
This document is an introduction to the PhD programme in the Department of Computer Science at University College London. It presents a guideline programme structure and covers many of the organizational issues of PhD study in the department.
The guidelines presented in this document are subject to change from time to time, as decided by the Department's Committee of Heads of Research Groups, and in consultation with research students.
All research student funding bodies assume that a full-time PhD programme lasts for 3 to 3.5 years. In particular, one can only usually get 3- to 3.5-year PhD studentships. Further the University of London has ratified a policy that 3 years is the standard period for a full-time PhD programme of study, with an additional year in ‘Completing Research Student’ status. The pressure to ensure that 4 years is the canonical completion period for full-time students is further re-enforced by various bodies (HEFC, CVCP, CPHC, EPSRC, etc.), which only take into account completions within 4 years in their assessments of departments. For part-time students there are corresponding expectations leading to an ‘on-time’ completion within 7 years. Pragmatically, however, 3 years of registered study and a few months of writing up seems the norm. It is Department policy that PhDs should be entirely completed within 3.5 years. In addition, the absolute maximum amount of time allowed by UCL from initial registration to submission of the thesis is 10 years, for both full-time and part-time students.
Given the above, it is important for the department to ensure that full-time students are able to, and do, complete within 3.5 years. This has two facets: Firstly, expectations of projects must be such that they are large enough to constitute PhD work but small enough to be finished within 3.5 years. Secondly, that the department provides support, both resource as well as supervision, to enable a 3.5-year completion. Support after 3.5 years is reduced. Support after 4 years should be zero.
Currently, UCL policy is that submission of a thesis must happen within the year (for full-time students) or the 18 months (for part-time students) following last registration, which is normally a 3-year period for full-time registered students. Separately there is the concept of a 'Course of Study', being the minimum period of registration before a thesis can be submitted. The minimum course of study is 2 years for full-time students and 3 years for part-time students. Since submission of the thesis can happen only after the completion of the 'Course of Study', departmental policy is to set this to the minimum to allow maximum flexibility. In particular, for full-time students the 'Course of Study' is usually set to 2.5 years allowing student to submit and be examined within the 3.5-year period.
This rest of this document sets out much of the information required by, and hopefully acts as a useful document for, students starting a PhD programme in this department. It is, however, a starting point, as there is much information you will need that is not here!
A PhD programme is many things:
- education in the area of interest and 'surrounding' areas;
- training in doing research;
- doing research;
- training in undertaking a lengthy project; and
- training in presenting ideas; learning to communicate ideas.
Fundamentally, a PhD programme is training how to think, how to construct hypotheses and arguments and how to communicate them, via papers, seminars and a thesis. Given that we are a Computer Science department, in the Engineering Sciences Faculty of UCL, the thesis itself needs to be on some aspect of computing.
Starting from the premise that a PhD is a 42-month activity, this section contains model timetables for full- and part-time students. These are not a rigid timetable but skeletons offered as a starting point for discussion between the student and supervisor to arrive at a timetable that allows interesting work to be done but admits a 3.5-year completion.
Registration -- initially MPhil registration.
General reading, directed by the supervisor, in the area of interest. This should be to get the student to the sharp end of the area and allow the student to appreciate what the research problems are. A few small, short projects are probably in order to vary the activities being undertaken. Seminars should be given on work done or on papers read
More detailed reading in, fairly narrow, more technical areas of a problem --- not a field since we do not wish to inhibit cross-disciplinary thinking. Aimed at becoming expert enough to tackle a thesis project. A smallish focused project is in order here to pin the reading on. Seminars should be given on work done and papers read. A report on the year's activities should begin to be prepared.
Months 8-9 FORMAL 1ST-YEAR VIVA
This is the first major examination, and must take place no more than 9 months from the start date. A feedback activity regarding the student’s ability to demonstrate mastery of the literature and state-of-the-art in an area of research. Given a read of the student's report, the supervisor, 2nd supervisor and an 'assessor' review the work done with the aim of providing the student with proper feedback on their work.
Detailed work on the thesis project. Give seminars on the thesis work. Towards the end of the period a report needs writing.
Months 21 FORMAL TRANSFER VIVA
This is the second major examination. A substantial project report is expected demonstrating the ability to conduct research, with initial research results, and a plan for completion of the work and writing of the thesis. The outcome of the viva will determine whether the student is allowed to transfer registration from MPhil to PhD. A failure to transfer normally results in the student working toward completion of an MPhil thesis. A second attempt at the viva must take place before month 24.
Thesis project work being tidied up and turned into a unified piece of work. Thesis writing being planned and chapters being drafted. Seminars being given.
Month 36 MOCK VIVA
A draft thesis and mock viva. This is to be attended by the supervisor, second supervisor and assessor and any others thought relevant. Thesis submission forms (aka Entry forms) completed and submitted.
Complete the writing of the thesis.
Clearly not all PhD programmes will take this long, and many are such that submission happens earlier than month 42, and even earlier than month 36, for example in month 32. This is to be encouraged, particularly if the student is working with a fixed term of funding (typically 36–42 months duration). The outline above is what we believe to be the longest that a programme should be. No PhD student should be expected to do more than 42 months work. If it transpires that an extra month or two is required to complete the writing up, then so be it. However, the goal is that this should be a rare situation.
What might go beyond month 42 is the examination of the thesis. The examiners of the thesis are normally expected to read the thesis with 1 to 3 months and to hold the viva examination very soon after that.
Part-time students vary significantly in their circumstances. Some may be working in UCL itself, even in this Department, whereas others may be working far away. Their outside work sometimes will cause periods when they are unable to stick to any rigid pattern of PhD work, and such periods should be discussed with the supervisor rather than hidden.
The basic norms of regular meetings with the supervisor, and also meetings with a second supervisor should be established early on in the course of part-time study. The basic pattern of progress as shown in the above section on full-time study should be followed, though the time-span is likely to be considerably longer. These are guidelines that should be discussed by supervisors and student at the start of the part-time PhD.
- 1st-Year Viva – Months 10-12 (determines continuation)
- Transfer Viva – Month 30 (and no later than Month 36, determines transfer to MPhil)
- Mock Viva and Draft Thesis – around Months 48–60 (determines submission status)
- Submission – around Months 60–72
Method of Assessment
Assessment has three purposes in the department. These are, in order of priority:
- To give feedback to the student on progress during the assessment period.
- To give the student practice at being assessed in the way that the final thesis will be assessed.
- To allow the department to monitor the progress of the student.
A number of variations on the theme of annual assessment have been tried by the department over the years with varying degrees of success. We have now settled on a procedure that involves each student writing a report and undergoing a viva at the end of months 9, 21-24 and 36 for full-time students.
For each viva, a report on the previous period's work written by the student, together with any papers written by the student during this period, is examined by a panel of three people; the supervisor, the second supervisor and an 'assessor'. The student then attends a viva with the panel. The panel provides a report paying particular attention to: the technical and professional development of the student; possible directions for further progress; and the relationship of the work to the spectrum of departmental research activities. The panel's reports are reviewed by the or the Research Student Tutor. At the same time a copy of the report is given to the student who has the opportunity to respond to any of the points made.
For the Month 9 viva, the student is expected to indicate an understanding of work underway and the issues to be pursued, preferably with a statement of the problem that the thesis will address.
In the Month 21-24 viva the student should be able to show substantial progress of the research and give a draft contents page for the thesis, indeed having this as part of this report is to be strongly encouraged. This viva normally results in the transfer from MPhil to PhD registration. At this viva, the student should be able to present a fairly concrete contents list for the thesis dissertation. Also they should have a plan for the writing of the thesis. Most of the actual thesis work should be complete but there will be tidying up needed and perhaps the odd extra piece of work to confirm (or deny) issues to be addressed in the thesis.
It should be noted here, that the department expects all students to be aware of other work that is being undertaken in the department and how it relates to their own. Students will be expected to provide evidence of this awareness at the viva.
Other Feedback Support
Students have a need for support in two critical areas. First of all, there is the need for feedback from a wide range of experienced researchers. This is seen as a prerequisite for staying on course and for knowing when enough work has been done prior to submission of the thesis. Presently, the main means of eliciting feedback are by:
- Talking to their supervisor.
- Writing Internal Notes.
- Giving research seminars.
The second critical area is that of receiving advice and guidance in the planning and management of both time and research. These are skills that can take many years to learn. Most research students do not have these skills, and they are critically reliant upon more experienced researchers to advise and guide them. For these reasons a great emphasis is placed on the feedback element of the end of year assessment procedure. This is also the reason why we strongly support the UCL Graduate School, which runs courses on Research Methods, Writing Theses and Papers, and Time Management and all research students are strongly advised to attend these courses early on in their study.
It is expected that the student and supervisor work together during the early part of the second year (and possibly up to the end of the second year) to produce the thesis proposal. In fact, in some circumstances the thesis proposal could be produced during the first year. In this case it would be revised for each following assessment report.
The proposal can be reinforced by specific work schedules, one submitted at the time of the proposal, and revised schedules included in the following end of year reports. These are planning tools; exercises to assure both student and supervisors that what is proposed is feasible.
It is important to appreciate that a thesis proposal is not a table of contents and thus has nothing to do with the department's criteria for transfer to PhD.
In addition to the timetabled assessments, it has been suggested by some students that there should be a review-on-request mechanism that allows student or supervisor to call for a review at appropriate points in the research. For the review, the student and supervisor would propose a panel to meet, say within the month, and review whatever the student has prepared. The student, supervisor or department could request that the panel prepare a report similar to the current timetabled assessment document.
The department does not have the resources to adopt such a procedure formally, although there is nothing to stop such a procedure happening if the supervisor and student so wish or for exceptional circumstances. The department does appreciate that the timetabled vivas arrangement can be insufficiently flexible at times. In some cases an accident of timing may result in a critical milestone being reached just a month or two after the assessment date.
It is expected that relatively few students will need to take advantage of this mechanism. It should not be supposed that the introduction of these sort of reviews lessens the importance of the timetabled assessments; these will remain the main assessment procedure in the department.
Transfer from MPhil to PhD
Students are initially registered for the MPhil degree. Unlike many universities, we do not automatically transfer students to PhD registration at the end of the first year, after a successful viva. Instead we have a policy of waiting towards the end of the second year, when the student can usually provide the basic structure of their thesis. If it is recommended that the student should make a second attempt at the viva, then this second attempt must take place before month 24. If the student is unable to pass the viva after a first or second attempt, the implication is that the panel do not feel that the student can achieve a successful PhD and should instead write up the work done so far and submit for an MPhil degree. It is expected that such a submission would happen by the end of month 26 or if necessary month 27. Such an occurrence is and should be very rare.
Once the recommendation to transfer from MPhil to PhD is made there is a very simple administrative procedure to follow which can be implemented in a few days at most. As the university regulations forbid submitting the thesis less than four months after entering for the examination, we would expect that all transfers will be agreed no later than six months before the expected submission date.
It must be emphasized that all students accepted as research students in the department are regarded as being capable of completing a PhD. If, after the passage of time, this no longer appears to be possible, then the department alerts the student to this fact and recommends the student to aim to complete for an MPhil. More often than not, this approach causes the student to re-appraise their work effort and approach, so that they end up submitting for a PhD with the support of the department.
The Structure of a Thesis
It is advisable for all students to look at some theses in their topic area early on in their studies to see examples of the goal of those studies.
The actual chapter structure of a thesis is dependent on the topic addressed and the personal style of the student. However, the following is a fairly generic structural guideline for a good thesis:
This should be a one page summary including the area of study, the questions addressed, the achievements of the work and the future directions. Its main purpose is to categorize the work, in particular so that examiners can be chosen.
This should be a short chapter 'setting the scene'. It should include a description of the area of study, the problems found in the area, the particular problems addressed in the thesis (the 'thesis statement'), why the question is important and how it relates to future work in the area, what the thesis achieves with respect to the thesis statement and the area. This section should finish with an annotated guide to the rest of the thesis, indicating how each part contributes to the questions set in the thesis statement. This part should challenge the reader to ensure that they understand the central thesis statement and ensure they want to read the rest of the work.
This sequence of chapters should include the following material:
- Background work in the area; a literature survey.
- Closely argued treatment of the problems in the area, highlighting those that will be addressed. The thesis statement with justification of its importance and connection to the area.
- Detailed plan how how the thesis statement will be 'proved'.
- Exposition of the work done.
- Detailed coverage of the results and achievements.
Summary of the thesis statement and results. Problems and/or incompleteness with the work. Future work to be done by others.
The regulations do not specify a maximum length of thesis for Computer Science but the range 80,000--100,000 words is a good guide for the maximum; this means about 300 pages maximum. Normally thesis are between 100 and 250 pages. A thesis must challenge the reader to read, whether or not the thesis statement is controversial. If the subject is controversial, it is important to ensure proper and detailed argument of the points. This, however, does not mean that non-controversial thesis statements admit sloppy argumentation or writing.
Addressing a controversial issue or obtaining a potentially controversial result never, in itself, leads to failure, the thesis is judged on its merit as a piece of scientific or engineering work. If examiners start trying to fail students simply because they disagree with the results, then we bring the full force of the departmental and college administration into action in defence of the student. Fortunately, this situation very rarely occurs.
One piece of advice though: Students should always invite their supervisor to be an observer at the viva. This is the surest way of avoiding the sort of situation indicated above.
Getting Ready to and Actually Submitting
Having transferred to PhD registration, and at least 4 months before submitting the thesis, a thesis entry form must be submitted. Once the entry form has been received by UCL Registry and Academic Services, the thesis itself can be submitted between 4 months and 18 months later. The expectation is that 5 months will be the average.
Please make a copy of the completed entry form prior to submitting it to Registry and Academic Services and give it to the Postgraduate Administrator.
At this point college administration swings into action. Within 2 weeks (usually), the supervisor is asked to suggest examiners. These suggestions are then sent to Registry and Academic Services who verify that the examiners appointed have the relevant experience.
Once the thesis is submitted, the supervisor is asked to contact the examiners and arrange a time for the viva. The examiners are free to choose the length of time they need to examine the thesis but 1 or 2 months is usually the time required. The viva must only happen after the examiners have completed their reading of the thesis.
After the viva, the student may be asked to make changes, usually within three months. In unfortunate circumstances, the student may be told to 'resubmit'. This means significant and major changes have to be made to the thesis and that thesis submitted.
Assuming no problems, and adding all this together, we have a 36–42 month programme followed by a, usually, 3 month period till the final pass and then a further 1 or 2 months to award of the degree. In most cases, the student can be working during the period of the examination of the thesis. Technically though the student must remain resident in this country between submission and viva. Overseas students are exempt from the rule but should remember that the viva will be held in London.
UCL regulations require all PhD students to submit their thesis within 10 years following their initial registration.
For information on what to do when you are ready to submit, see the CS guide on thesis submission in the Adminstrative section.
The Role of Key Staff
The department operates a two-supervisor system, thus every student has a first and second supervisor. The first supervisor is the main supervisor.
The initial matching of student to supervisor is made on the basis of the research interests of the applicant and the availability of a suitable member of staff with the time to take on the commitment. Via an interview, where this is feasible, the applicant and putative supervisor are able to determine whether or not they are matched technically and temperamentally. If, though this is rarely the case, a mismatch does occur and change is desirable, the Research Student Tutor discusses the situation with the current and potential future supervisor to ensure a smooth and amicable transfer.
One constraint we do have on supervisor changes is that the basic area of research should not change once students have completed their first year. After this period, we only sanction transfer for relationship reasons, not because the student would prefer a change of topic.
The student/supervisor relationship is a critical factor in PhD research. To date we have used the EPSRC document "Research Student and Supervisor" as our guideline for the conduct of these relationships, although it should be emphasized that each student/supervisor will normally, and most effectively, adopt a mode of working that suits them jointly. Having said this, the department continues to monitor progress. The Research Student Tutor, in conjunction with all academic staff, continues to actively seek ways of ensuring that supervisors encourage project work that admits of a three year completion.
At all times, the supervisor acts as an adviser to the student. PhD students are not employees who are given tasks to be completed in 2 weeks. The student is undertaking some work of their own with the supervisor providing constructive criticism, advice, encouragement etc. The relationship between the supervisor and student is critical to the success of the PhD programme, it must be one based on trust and friendship not based on a master--slave relationship.
Throughout the period of study the student is expected to meet regularly with the first supervisor, once per week minimum.
The practice of assigning a second supervisor was prompted by three things; pressure from the students, the knowledge of the 'student panel' system that operates in the USA and also a recommendation by the UCL Academic Development Sub-Committee that "the system of appointing a 'second supervisor' for each research student (in order to widen the range of committed advice available to him or her) should be adopted throughout the College as far as possible", which itself was prompted by statements emanating from the research councils.
As research students perceive a genuine need for a committed second opinion to be available to them on a regular basis, the second supervisor is chosen on both technical and personal grounds and is asked to make a definite commitment to maintain communication. The second supervisor is expected to review, criticise and comment on written work, but not necessarily to keep fully abreast of the student's work as it progresses, which is the role of the first supervisor. We expect a staff member in the role of second supervisor to spend of the order of one or two hours per month on supervisory activities, in addition to reviewing written work.
It is vital that the second supervisor is not merely a clone of the first. The second supervisor must be an individual who can bring different skills and qualities to the role, who is able to take a detached view of the first supervisor/student relationship, but has a good interest in the research area.
We expect each student to meet the second supervisor for one to two hours per month.
The first and second supervisors' main responsibility is the academic guidance of research students. They should also be responsible for much of the pastoral care of students. However, there are a number of other issues, mainly pastoral and administrative, that come within the remit of the Research Student Tutor. Briefly, the Research Student Tutor is responsible to the Head of Department for the implementation of departmental policies as they affect the body of research students as a whole and for sundry duties related to the well-being of individual students. In connection with this role the Research Student Tutor liaises mainly with the Research Student Representative although individual students with particular problems are encouraged to see the Research Student Tutor in person.
Typical matters dealt with by the Research Student Tutor over the last year or so are: negotiating changes in supervision, advising on matters associated with registration and examination, organizing the assessments, and responding to the requests of financial sponsors for progress reports.