I've just started a new discussion thread titled "Your experience with the class",
I have copied my post below. If you would like to add to this thread it would be much
appreciated!

Cheers,
Mike

=========================================================================================
I wanted to start this thread where people could post about their experiences with the class - what their background is, what they think of the material, why they took the class, what they got out of the class, and so on. It would be very useful for Alexander and me to get an idea of what people got out of the class.

I also wanted to say how much I've enjoyed teaching this class! The discussion boards are the main way that we get to know the students, and I have been constantly amazed by how insightful people's comments have been, and how helpful and constructive students have been when working with each other. Sincere thanks for making this such a positive experience :-)

Mike

Fri 3 May 2013 7:43 PM CEST

We have now posted PA4 on global linear models. While the assignment is optional, we think it will be helpful for gaining practical experience with the ideas presented in the last few lectures.

The assignment walks you through building a GLM for tagging from scratch. With a little extra work, you can build a tagger that scores significantly better than the HMM from PA1.

Good luck,
Alex

Thu 2 May 2013 6:05 PM CEST

Week 10 lectures have now been released!!

This week we're covering global linear models for two problems: first tagging, and second, dependency
parsing.

Cheers,
Mike

Mon 29 Apr 2013 3:18 AM CEST

The reference solutions for PA1 and PA2 are now posted in the forum. You can access them at PA1 solution and PA2 solution. We hope they are useful for future coding.

Cheers,
Alex

Thu 25 Apr 2013 4:48 AM CEST

We have now released quiz 4 on the site. This quiz covers weeks 7 and 8 of the class on log-linear models.

Good luck,
Alex

Mon 22 Apr 2013 11:32 PM CEST

This week we're covering: 1) the Brown clustering algorithm (a method for unsupervised
learning of word clusters from large quantities of unlabeled data); 2) Global linear models,
and the perceptron algorithm.

Cheers,
Mike

Sun 21 Apr 2013 7:52 PM CEST

Programming assignment 3 on translation alignment is now posted. It is due on May 1st.

We highly encourage students to make use of the discussion forum while working on the assignment. Posters have been great about helping fellow students find bugs and optimize their code for the previous assignments.

Good luck,
Alex

Mon 15 Apr 2013 7:32 PM CEST

This week we're covering 1) log-linear models for tagging (sometimes referred
to as "MEMMs"); 2) log-linear models for history-based parsing.

Cheers! Mike

Mon 15 Apr 2013 2:42 AM CEST

I have now released week 7 lectures, and quiz 3, on the site.

In week 7 we are starting a new topic, log-linear models.
These models are widely used in modern NLP systems.

Quiz 3 covers weeks 5 and 6 of the class (the material
on machine translation).

Cheers!
Mike

Mon 8 Apr 2013 2:49 AM CEST

Week 6 of Natural Language Processing starts today -
lectures have now been posted. This week covers phrase-based
models for statistical machine translation.

Cheers,
Mike

Mon 1 Apr 2013 2:03 AM CEST

Programming assignment 2 is now posted. The assignment covers PCFGs and parsing. It is due on April 15th.

We highly encourage students to make use of the discussion forum while working on the assignment. The discussion about PA 1 has been very constructive, and students have been great about helping others avoid potential points of confusion.

Good luck,
Alex

Wed 27 Mar 2013 6:57 PM CET

Week 5 of Natural Language Processing starts today, and the lectures have now been posted on the website.
In weeks 5 and 6 of the class we'll be covering a new topic, machine translation.

As usual, notes accompanying the lectures are at

http://www.cs.columbia.edu/~mcollins/notes-spring2013.html

Cheers,
Mike

Mon 25 Mar 2013 1:56 AM CET

For various reasons we've decided to give an extra week for programming assignment 1 -
it is now due on April 1st (a week from this coming Monday).

Cheers, Mike

Fri 22 Mar 2013 6:42 PM CET

Quiz 2 has now been posted. It is due on April 1st, and covers material from weeks
3 and 4 of the class.

Cheers,
Mike

Thu 21 Mar 2013 1:35 AM CET

The solutions to quiz 1 have now been posted at
this link.

Cheers, Mike

Tue 19 Mar 2013 5:07 PM CET

Tomorrow is the start of week 4 of the course! A few announcements:

1) Week 4 lectures have now been released (sorry that close captions aren't yet
available for all segments - closed captioning has been requested though, so should
be up soon).

2) Quiz 2, covering weeks 3 and 4, will be released in the next few days. It will be
due on April 1st. (Recall that Quiz 1 is due on March 18th, i.e., this coming Monday.)

3) On the grading policy for the class: 50% of the final grade will be from Quizes
1-5; 50% of the final grade will be from the three programming assignments.
Quizes 1/2/3/4/5 will cover material from weeks 1-2/3-4/5-6/7-8/9-10 respectively.

Cheers,
Mike

Sun 17 Mar 2013 1:51 AM CET

A couple of updates:

1) Quiz 1, which covers material from weeks 1 and 2 of the course, has
now been released. The due date is March 18th.

2) Week 3 lectures have now been released (in case you would like
to get ahead on the course - yesterday was the start of week 2 of
the course).

Cheers, Mike

Mon 4 Mar 2013 10:22 PM CET

The first quiz will be due March 18th, the first programming assignment will
be due March 25th. Apologies for the error in the previous email/announcement.

Thu 28 Feb 2013 9:30 PM CET

A few people have been asking about quizzes and programming assignments for the course - sorry
for the delay in announcing this.

The first quiz for the course will be released in the next couple of days. It will cover material from
weeks 1 and 2 of the course - it will consist of multiple choice questions. It will be due on March 18th
(a week after week 2 finishes).

The first programming assignment for the course will also be released in the next couple of days.
In this assignment you will implement a trigram HMM tagger (as described in week 2 of the course).
This assignment will be due on March 25th (2 weeks after week 2 of the course ends).

Thanks,
Mike

Thu 28 Feb 2013 9:02 PM CET

Just a reminder that the NLP course on Coursera starts tomorrow (Sunday 24th)!
The first two weeks of lectures for the class have now been posted.

I will be posting comprehensive notes for the class - similar
in style to an online textbook - at


http://www.cs.columbia.edu/~mcollins/notes-spring2013.html

I'd strongly recommend reading the notes. They follow the video
lectures closely.

Cheers,
Mike

Sat 23 Feb 2013 11:34 PM CET

For those of you who have already watched the lectures on the Viterbi
algorithm in week 2 of the course, there was a small but important
bug: see this thread for where it was pointed out and where I've also posted the correction.

We've now updated the video lectures to have the correct version of the algorithm. You can also see the hmm notes for a description with a correct version of the algorithm.

Sorry for any inconvenience with this! The first programming assignment
uses the Viterbi algorithm so we wanted to make sure to get it completely
correct in the videos and notes.

Thanks,
Mike

Thu 21 Feb 2013 5:23 AM CET

Welcome to Natural Language Processing (NLP). NLP is, I think, a tremendously exciting field. It draws on insights from computer science, linguistics, and as we'll see, increasingly from statistics and machine learning. NLP is having a huge impact in our daily lives: many applications and technologies are now making use of basic ideas from NLP. In this class you will learn about state-of-the-art methods for a broad set of NLP problems and applications.

Fri 8 Feb 2013 6:43 PM CET