Scientific Project Computer Engineering (182.759)

[TISS-Seite] [Syllabus] [Aims and Scope] [Enrolling] [Grading] [Schedule] [The Project] [Other resources]


Aim

This course permits first steps in own scientific work on selected topics of computer engineering. Based upon a (typically self-assigned) publications on some suitable topic in this field, a short paper shall be written and presented during the semester. In addition, the course is also an excellent opportunity to get in touch with our research projects and establishes a solid basis for related diploma thesises and dissertations.



Subject

Possible topics (selection): Hybrid failure models, partially synchronous system models, self-stabilizing distributed algorithms; real-time scheduling, topology control and routing in dependable wireless networks; diverse limitations of the synchronous design paradigm, metastability and synchronizers, asynchronous design styles, delay models, testing and fault tolerance for asynchronous logic, self-healing circuits. An overview of current computer engineering research topics and potential supervisors at the contributing institutes/working groups is provided in the TI research presentations (http://ti.tuwien.ac.at/teaching/ti-research-presentations).



Lecturer

Ulrich Schmid, Muhammad Shafique, Andreas Steininger

Homepage

http://ti.tuwien.ac.at/ecs/teaching/courses/sciproj/

 



AIMS AND SCOPE

This is a graduate-level optional course that aims at first steps in own scientific work in computer engineering. The goal is to find, select and read a few (theoretical) papers related to some individually assigned topic in the field, and to write and present a short LaTeX paper that unifies/integrates/extends the results in some way; exceptional papers can and shall be submitted to a conference or a journal. More details are available below.

Prerequisites are basic knowledge in at least two areas of computer engineering (in particular, fault-tolerant distributed algorithms, real-time scheduling, asynchronous digital design, dependable VLSI circuits) and basic knowledge and interest in scientific work (level of 180.765 Wissenschaftliches Arbeiten).

 


 ENROLLING

There is no need to enroll into this course, but you do need to find a supervisor at our institute who is able and willing to take your topic.


GRADING

Grading will be based on the following components:

  • 1st paper assessment (20%): Assessment of the first version (= "submitted version") of your paper. In order to get some early feedback, you also have to present your paper; note that this first presentation is not graded.
  • Reviews (10%): Quality of the reviews of the assigned papers (if any).
  • Paper presentation (40%): Assessment of the performance in presenting the paper, according to usual (conference-type) criterions.
  • 2nd paper assessment (30%): Assessment of the final version of the paper.

In addition, attendance of at least two TI Research Presentations is mandatory.

Outstanding work may be submitted to a regular scientific conference in the field. In case of acceptance, the expenses for attending this conference will be paid!

 


 SCHEDULE

The schedule is maintained dynamically. All presentations will be announced under TI Research Presentations  and usually take place in the library of the Embedded Computing Systems Group (E182/2), Treitlstraße 3, 2nd floor.

 


THE PROJECT

 

Purpose: The project assignment has several purposes, namely,

  • introduction to the scientific literature in the area,
  • application of ideas and techniques presented in basic courses in a more extensive and open-ended environment than in homeworks,
  • practice in writing technical documents (in LaTeX),
  • practice in reviewing scientific papers,
  • practice in presenting scientific papers.

Assignment: Your project will be based upon some topic chosen by yourself (subject to the approval of your supervisor, of course). 

Selection: After acquiring (from the library or the internet, see the tips below) and reading your "starting" paper(s), you have to find, select and read a few related paper from the literature. Some amount of work should go into the selection process. Your choice should focus on paper that have a significant theoretical component. An implementation-related topic is possible, however, if it relates somehow to theory, for instance, a simulation to verify or discover the average case performance of some algorithm. Some general ideas for how to select a suitable paper:

  1. Read a few related theoretical papers and choose one (maybe even two) that allows to either
    • propose a simpler version of a problem presented in a paper and develop a simpler solution to your problem. OR
    • discover a new connection between the papers; for instance, show how the results in one paper can be used to improve or simplify the results in another paper. OR
    • develop a new notation and/or result that can simplify and/or unify the results in these papers and redo the results in your new notation. OR
    • solve an open problem related to the papers.
  2. Come up with a new problem based on some application related to the "starting" paper and try to solve it.
  3. Study the scientific basis of a commercial product that is based on the "starting" paper.

 

Paper: The paper should summarize and critically review the selected work w.r.t. the "starting" paper(s), and either extend it in some way or simplify the results by making some simplifying assumptions. It must be self-contained in that ot should not be needed to consult the original papers for general understanding. The paper is to be typed in LaTeX using the IEEE conference style, and should be at most 10 pages long; if you target some specific conference, you can also use the particular style used there. Part of your grade will be based on the quality of your composition (including spelling, grammar, logical flow). The typical organization for a technical paper in this area is:

  • introduction (informal explanation of problem, why it is important/interesting, overview of the paper contents)
  • formal definitions
  • results, including intuitive explanations
  • conclusion (including open problems).

As a general rule, very little, if any, of your paper should consist of copying the contents of the papers you read, and of course if you do quote from a paper, be sure to credit the source.

[To understand the rationale of the above, it is very instructive to see how rigidly top US universities handle the issue of academic integrity, plagiarism, etc. Consult the Texas A&M University Code of Honors for an example.]

Reviewing: You will be assigned some paper(s) for reviewing. Please be sure to adhere to some reasonable reviewing standards, as e.g. known from the seminar 180.765 Wissenschaftliches Arbeiten. Note that the quality and appropriateness of your review will affect your grade, so please refrain from unduly praising a bad paper [or unduly rejecting a good paper].

Milestones: The milestones need not be synchronized with your colleagues but can rather be individually chosen, even across semester boundaries.

  • Topic assignment: Choose your "starting" paper (please make sure to also send me the .pdf in case of a choice of your own). Note that a single topic and/or "starting" paper may be chosen by several participants, but you need to convince me that there are sufficiently many disjoint project proposals for the same topic.
  • Project proposal due: Turn in your project proposal. The proposal must consist of half a page describing what you plan to do, as well as the .pdf of your selected papers. Please note that if your plan is not sufficiently convincing, your supervisor will not approve it, so it is best to talk with him in advance about your plans, in particular, if you are considering borderline topics
  • First presentation: Make a first conference-style presentation of your intended paper (20-25 min. talk + 5-10 min. questions) to get some early feedback. Of course, you are free to ask other colleagues to help you improving your paper (e.g. proof-reading) before turning it in.
  • Paper due: Turn in your paper. [This step corresponds to the submission of your paper to a conference.]
  • Reviews due: Turn in your reviews (if any).
  • Presentation : Make a conference-style presentation of your paper (20-25 min. talk + 5-10 min. questions).

OTHER RESOURCES

 

A few links for searching literature:


 Miscellaneous: