Expose

How to Expose #

Prior #

  • You got some topic ideas from me. You can pick one to reserve it for 7 days to decide if you want this topic.
  • After the 7 days the topic is considered open again and I might give it to another student.

Formals Expose #

  • The expose is mandatory.
  • Everything written in the Expose can later be directly transfered to the Thesis. You are not wasting your time.
  • You picked decided on a topic and told me? Get the template from https://www.inf.uni-hamburg.de/inst/ab/snp/courses/material/templates.html which includes every formality.
  • You now have around 6 weeks to finish your expose.
  • The following times are suggestions..
    • Sent me a first, raw version in the first 2 weeks.
    • Sent me another almost complete version in the first 4 weeks.
      • This version should include all content points (see below) but not necessarily in full detail.
    • Sent another version to the first supervisor within the 6 weeks.
  • Continue writing even if you are waiting for feedback, I try to be timely on this
  • You have the ok of the first supervisor and me?
    • Remind me to add an entry in the thesis tool.

Content Expose #

  • Introduction (0.5-1 page)

    • Introduction to the topic and the problem you are trying to address. Why this topic (motivation)? Why do we need a new approach/solution for the topic? For more information see Thesis#Introduction

  • Requirements and Related work (4-5 Pages)

    • Requirements (~1 page)
      • Continue with Related Work first, come back later.
      • Formulate requirements for your solution (low false positives, fast, privacy, …)
    • Related Work (3-4 Pages)
      • The most important section for you in this moment. What is already there? Where to improve? You really want to get an overview of the field. Based on this knowledge you will pick your approach.
      • What is the state of the art?
      • What is not covered in the state of the art that you want to cover?
        • e.g., related work is not scalable, privacy friendly, …

  • Method (2-4 pages)

    • How do you plan to approach/solve the problem?
      • Proof of concept
        • Machine Learning, Crypto, …
        • If you do something with Machine Learning you need to describe how you get the necessary data. Best: Find a dataset that can be used. Second best: Create the dataset on your own. Okay: Merge two or more datasets or process them. Okayisch: Generate your data by yourself.
      • Theoretical proof
      • Empirical study

  • Evaluation (1-2 pages)

    • How to evaluate?
      • This does not need to be complete, but you should think about ways to evaluate. Often you could build some software and evaluate its performance.
      • The best way is to find similar literature and see how they evaluate it and then recreate it.
    • What are your expected results?

  • Time plan. Do this when you are done with everything else. (~1 page)

    • High level overview, plan on two weeks granularity.
      • Keep in mind that the Expose will take some time.
    • Give yourself more time that you would think.
    • Aim to be done one month before the deadline.
      • so you have one month as backup. You will likely need it.