Final Project

Overview

Your final project involves a mixture of theory and implementation. You have quite a lot of freedom to select a topic. One direction would be to take some of the ideas and components that were discussed in class and embed them in a mini-system that actually performs a task that people might be interested in. Nobody wants to know about tokenization and tagging, but people may be interested in seeing a summary of a document, a list of named entities that are becoming more prominent, or a summary of people who are most likely to start blowing up things.

We provide two example topics: entity chronicles and conditional random fields for natural language processing. Follow the link above for some more thoughts on the entity chronicler, contact Ben Wellner (wellner@cs) for more ideas on how to use conditional random fields.

As mentioned before, you are free to select your own topic and we will be lenient as to what topics are allowed and how strongly you need to motivate them. For example, you do not need to prove that the program you will write is useful to non-linguists. But topics do need to be a bit more general than "improving my tokenizer".

Some rules:

  1. You are allowed and encouraged to work in groups.
  2. You are also allowed to combine this with a project for another class, but this won't save you any time, the requirements will change accordingly.
  3. Borrow, re-use and quote at will from others, but praise (or blame) the sources.
  4. Again, use Python or get permission to use another language.

Timeline

First create your groups and select a topic. You can contact us any time for ideas and feedback. By March 31st, we'll need a 1-2 page write-up saying:
  1. what you're doing
  2. who's doing what
  3. programming language you want to use
We'll contact you and set up a little meeting.

Requirements

Due date is May 1st.

Your work should include

  1. discussion of the issues involved, include a small literature study.
  2. your program

Presentation

Prepare a 10 minute presentation for April 26th or May 1st.

The day you present does not effect your due date.