There is a website at www.knollop.com that has evaluations of online courses, including all Coursera courses. If you want to contribute an evaluation of this course, go to:

http://www.knollop.com/courses/3145/Neural-Networks-for-Machine-Learning/

Fri 21 Dec 2012 1:23 PM PST

Dear students,

The certificates have been issued to people who got 70% or more overall. You can view and print them here. They don't show your grade, but you can calculate it using this spreadsheet.

We hope that you enjoyed the course, and we want to thank you for being such a nice student community. Before you go, please fill in the student survey, if you haven't already.

Congratulations on finishing the course, and happy holidays!

Geoffrey Hinton, Tijmen Tieleman and the rest of the staff

Fri 21 Dec 2012 12:09 PM PST

I have just uploaded the videos and slides for lecture 16.
This material is all optional: its not on the final test.

It should be visible soon.
Sorry its so late.

Geoff

Mon 26 Nov 2012 4:28 PM PST

Please take a few minutes, at some point, to answer some basic questions like "how much time did you spend on the course" and the like. The survey is posted here. The Coursera website seems to think that it's a quiz, but of course it isn't. Apologies for that confusion.

Thank you!

Tijmen

Mon 26 Nov 2012 10:34 AM PST

The final test is now on the web and is due in 8 days time. There will not be any extensions to this deadline, and it is a hard deadline: if you submit your answers after the deadline, you will have a score of zero.

Please note that you are only allowed one attempt: It's a real test.

We have done our best to avoid really hard questions. If you managed to do the programming assignments we think you will not find this test too difficult.

Good luck and congratulations on getting this far.

Sun 25 Nov 2012 9:01 PM PST

For the final week, there is only one weekly quiz (for lecture 15). Lecture 16 will be a brief overview of all the important topics that I could not fit into the course. The material in lecture 16 will not have a weekly quiz and it will not appear on the final test. Also, I have not managed to record it yet.

On Monday Nov 26, the final test will be posted. It is worth 25% of your final grade. You will have a week to complete it and, unlike the weekly quizzes, you will only get one chance. I will send another email about it with more details next Sunday.

Sun 18 Nov 2012 9:01 PM PST

The deadline for programming assignment 3, which was today, has been extended by three days.

By the way, if you're struggling with the math, you'll find this forum post very helpful.

Tijmen

Tue 13 Nov 2012 9:41 AM PST

For programming assignment 3, there's a lively forum discussion going on, with lots of valuable questions and insights. If you feel that you're stuck with PA3, then I suggest that you do some reading (and maybe talking) on the forum. Also, I've published several improved versions of the gradient checker. Those are mentioned in announcements that I add at the top of the assignment description. Please check for news there, every once in a while.

Tijmen

Mon 5 Nov 2012 4:39 PM PST

Quite a few people lost some days due to the big storm on the East coast of the US. I have therefore extended the deadline for quizzes 7 and 8 by 2 days. If you lost more days than that you should use some of your late days. That is one of the things they are for.

Geoff

Sat 3 Nov 2012 3:22 PM PDT

I would like to thank the 99.97% of you who have realized that I don't have time to answer email and prepare the videos. It a very impressive percentage and I really appreciate it.

Geoff

Wed 31 Oct 2012 9:01 AM PDT

Dear students,

Programming assignment 3 is now available. Good luck!

Cheers,

Tijmen

Tue 30 Oct 2012 12:01 AM PDT

This week's videos are about ways of making neural networks generalize better. They explain a lot of the tricks that have been used to make neural networks more useful in practical applications. It was hard to fit in all of the techniques I wanted to describe so the videos are a bit longer than in previous weeks.

The big storm that is hitting the North East of the US may affect Coursera's servers so I have extended the deadlines for weekly quizzes 5 and 6 by a few days and also extended the deadline for programming assignment 2 by a few days. This is intended to allow for up to 2 days of downtime, so please don't ask for extra extensions unless there is more downtime than that. I am not going to extend the deadline for programming assignment 1 any further because the website has now been up reliably for several days past the original deadline.

The third programming assignment is a day behind schedule. I will try to get it on the website by Monday night (Eastern Standard Time).

Geoff Hinton

Sun 28 Oct 2012 6:34 PM PDT

The Coursera website was down for part of last Monday when the first programming assignment was due. We are therefore extending the deadline to next monday at 11.59pm EST (and the final deadline will be extended to a week after that).

Some people used up a late day to ensure their submission was not treated as late, so we have increased the number of allowed late days from 7 to 8.

Thu 25 Oct 2012 1:00 PM PDT

We are going to try to change the penalty for quizzes and assignments that are submitted after the first deadline but before the final deadline. Currently this penalty is 50% and we are going to reduce it to 25% (so you would get 75% of the marks if you got all of the questions right). We believe the system can automatically regrade quizzes and assignments that have already been graded. We will let you know soon if this doesn't work.

Thu 25 Oct 2012 11:38 AM PDT

There is a visualization technique called t-sne that will layout the word vectors in a two-dimensional map, with very similar vectors very close to each other. This lets you see which words it thinks have similar meanings. Using T-SNE is not part of the assignment, but its fun.

T-SNE visualization software can be downloaded from
tsne.tar.gz
OR
tsne.zip
Take a look at README_tsne.txt for using it.

For more details about T-SNE see this.
Below is an example of the kind of embedding you would get

Fri 19 Oct 2012 9:01 PM PDT

I apologize. The email I got from MathWorks did not mean that MATLAB was free for people taking the course. It just meant that they would sell it to you at the student rate rather than the full rate.

Geoff

Mon 15 Oct 2012 12:44 PM PDT

Despite what I say in Lecture 5b, there is no lecture 5e. By the time I finished lecture 5d, there was already sufficient material so I decided to talk about the last method for achieving viewpoint invariance in lecture 16.

Mon 15 Oct 2012 11:30 AM PDT

If you are taking my course, you can now get a free copy of MATLAB, which should be easier to work with than Octave. You will have to interact directly with MathWorks as described below. Please do not ask me or the TA's for any help on this. The information given below is all we know. We do NOT guarantee that the Octave code and data we have supplied for this week's programming assignment will work unchanged in the version of MATLAB they supply, but any changes required should be minor.

I received email from Jerry Brusher who works in Education Technical Marketing for MathWorks (makers of MATLAB and Simulink). Here is an edited version of his email:

Regarding your Neural Networks for Machine Learning course on Coursera, I wanted to let you know that MathWorks has elected to make MATLAB and Simulink Student Version available to individuals taking classes through Coursera. Simply have your students list Coursera as their institution and submit a screenshot of their registration page as proof of student status. We would be very grateful if you could pass this information along to your students.

Jerry Brusher, Ph.D.
Education Technical Marketing, MathWorks
Jerry.Brusher@mathworks.com

Mon 15 Oct 2012 10:59 AM PDT

On Monday Oct 8, you will be able to see lectures 3 and 4 of my Neural Networks course and the two weekly quizzes that go with those lectures. You will also be able to see the first programming assignment which should be quick and easy, but its not at all obvious HOW you see the assignment.

To see the assignment, you have to start doing a Quiz called "Programming Assignment 1". When you start on this quiz there will be a long preamble that tells you where to get code and what to do. You can then leave the quiz without submitting your answers yet. Once you have got the code working you will be able to answer the questions in the quiz and you will be allowed to submit two times and get the maximum of your two scores (as with the weekly quizzes).

The main point of this assignment is to get you used to using Octave.
There are some nice videos on learning to use Octave in Andrew Ng's machine learning course. You can view these videos without signing up for his course by going to https://class.coursera.org/ml/lecture/preview/26

Sun 7 Oct 2012 9:01 PM PDT

The quiz for lecture 3 will be available at the correct time, but the quiz for lecture 4 will be delayed by up to one day.

Sun 7 Oct 2012 9:01 PM PDT

The version that got captioned was the unedited version.

I will try to replace it with the edited version. It wont have captions but they should arrive in a few days.

Sun 7 Oct 2012 9:01 PM PDT

Many of you are unhappy with only being allowed to attempt a quiz once. Starting in week two, we have therefore decided to make up twice as many questions and to allow you to do each quiz twice if you want to. The second time you try it the questions will all be different. Your score will be the maximum of your two scores. For week one, the quizzes will remain as they are now.

Many of you would like the names of the videos to be more informative. We will change the names to indicate the content and the duration.

Some of you thought that some of the quiz questions were too vague. We will try to make future questions less vague.

Some of you are unhappy that we do not have the resources to support Python for the programming assignments. We sympathize with you and would do it if we could. You are still welcome to use Python (or any other language) if you can port the octave starter code to your preferred language. We have no objection to people sharing the ported versions of the starter code (but only the starter code!). However, if you get starter code in another language from someone else, you are responsible for making sure it does not contain bugs.

Tue 2 Oct 2012 11:10 AM PDT

The videos

Each week, the course will have two lectures. Each lecture will be divided into about five short videos. Both lectures will become available on Monday (at 12.01 am Eastern Standard Time). There are quiz questions embedded in the videos to help you check that you are understanding the material. These quizzes do not count towards your final grade.

The two weekly quizzes

Each week, there will be two quizzes (one per lecture) that do count towards your final grade.
The two weekly quizzes will become available at the same time as the two lectures. To help keep you on track with the course, we will have due dates associated with the weekly quizzes. After the due dates, we will continue to accept late submissions with a penalty on your score.

The programming assignments

Programming assignments will become available at the beginning of weeks 2, 3, 5 & 7. You will need to answer questions about the results produced by the programs and your answers will count towards your final grade. As with the weekly quizzes, there will be due dates for the programming assignments.

The first programming assignment will be very simple and will count less towards your grade. It is mainly intended to get you to download Octave and get used to using it (see the Octave installation link). We regret that we do not have the resources to support other languages, but if you have Matlab it should be simple to adapt the Octave code we provide. You will not need to submit any code.

The final test

On the monday of the ninth week the final test will become available and should be completed by the following monday.
The final test will be 25% of the final grade, the programming assignments will be 35% and the weekly quizzes 40%.

Sun 30 Sep 2012 9:01 PM PDT