Last updated: 2/23 (Mon) 5:45 PM PST. Details here.
Due Date: 2/25 (Sat) 11:59 PM PST.
Hard deadline: 3/1 (Wed) 11:59 PM PST with 4 late days. Assignments turned in past this date will be penalized (see the grading page).

In this assignment you will learn about named entity recognition and implement a baseline window-based model, as well as a recurrent neural network model. The assignment also covers gated recurrent units, applying them to simple 1D sequences and the named entity recognition.

Setup

Note: Please be sure you have Python 2.7.x installed on your system. The following instructions should work on Mac or Linux. If you have any trouble getting set up, please come to office hours and the TAs will be happy to help.

Get the code: Download the starter code here and the assignment handout here.

Python package requirements: The core requirements for this assignment are

  • tensorflow
  • numpy
  • matplotlib

If you have a recent linux (Ubuntu 14.04 and later) install or Mac OS X, the default TensorFlow installation directions will work well for you. If not, we recommed that you work on the corn clusters. TensorFlow is already installed for the system default python on corn.

We will also be providing access to GPU computing facilities on Microsoft Azure. Details on how to access these facilities will be uploaded soon.

Submitting your work

Do not code outside of the "# YOUR CODE HERE" blocks, modify the list of imports, change function names, etc. (tuning parameters is fine). Make sure your code runs before submitting. Crashing due to undefined variables, missing imports, hard-coded dimensions, and bad indentation will lead to significant deductions.

For the written component, please upload a PDF file of your solutions to Gradescope as a group (i.e. one submission per group). Remember to add you group members before submitting. When asked to map question parts to your PDF, please map the parts accordingly as courtesy to your TAs. This is crucial so that we can provide accurate feedback..

To submit your code, please see the following Piazza post.

Assignment Overview (Tasks)

There will be three parts to this assignment. Question 3(d) depends on 2(c), but otherwise the rest of the assignment can be done independently. We recommend reading the assignment carefully and starting early as some parts may take significant time.

Q1: A window into NER (30 points)

Q2: Recurrent neural nets for NER (40 points)

Q3: Grooving with GRUs (30 points)