Machine Learning and Natural Language

Spring 2009

Critical Reading Essays (updated 02/10/09)

The additional reading material in this class was chosen to complement the material presented in class. Most papers chosen augment the lectures by describing applications in which the ideas/techniques are used, or investigate some theoretical ideas more deeply. This is by no means an exhaustive collection of the major papers in the area, but rather as a relatively representative papers that can collectively serve as a good starting point for research on learning techniques in natural language processing.

Instructions:

There will be 4 critical reading essays.
Though I cannot announce deadlines for all the critical reviews now, current deadlines are
Please send them to me by email whenever you want. Make sure you send all four essays before the beginning of the final project.

If you do not give a presentation in class (did not volunteer in time) you will need to submit 6 of these.

You can choose any of the papers listed under additional Readings or student presentation between the time of the last submission and the current one.

If you choose a paper that is presented in class you must send the critical essay before it is presented.
You are encouraged to read more papers from the list, and follow up on the references in the papers listed (in some cases, you will have to do it in order to understand the paper).
In addition to get you better acquainted with key ideas and applications, the goal is to train you in read a research paper both quickly and effectively.
Your essay should not exceed 2 pages (1 is ideal). Spend only 1-2 paragraph introducing the problem and the approach taken. The rest of your writing should be devoted to critic, where you can address any issue you feel is important in the context of this paper. Examples might be: (1) assumptions made in the paper; (2) evaluation; (3) generality of the proposed method (is it applicable for other problems) etc. For the list of additional readings follow the course plan page