ELEC-E5550 - Statistical Natural Language Processing D, Lecture, 9.1.2024-16.4.2024

Writing a literature review and project plan

You can start the submission PDF with (an up-to-date version of) your team abstract, for context.

Literature review

The goal of the literature review is to get to know the topic of your project. The literature review should cover the broader topic (e.g. word embeddings spaces and analogies) and the specific methods that you plan to use (e.g. bag-of-words word2vec).

Show what the practical or theoretical difficulties are in the broader problem. Show the general framework of how the problem is typically approached (sometimes it is relevant to briefly describe a couple of broad strategies). Show how results in this topic are typically evaluated. Show how the specific methods that you apply relate to the larger topic at hand. You may also find it interesting to briefly showcase some recent state of the art results.

You should follow academic citation principles: back specific claims with specific citations. A typical length of a literature survey could be 2 pages: prioritize quality. Assume that your topic is totally new to your student peers, and for such an audience.

Project plan

After the literature review, write a less formal project plan: describe what you intend to do, with what data, and how you will evaluate. Again, write for your student peer audience, try to be clear and specific enough that they get a general idea of what you are going to do in practice.

A good project plan should have some flexibility - have simpler goals, and more extended ideas that you can reach for if you have time.