Stanford University: Probabilistic Graphical Models

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Stanford University: Probabilistic Graphical Models

pathematica
This post was updated on .
Further to the recent free course on machine learning (presented by Prof Andrew Ng), as you may know, Stanford are offering another free course entitled "Probabilistic Graphical Models" (presented by Prof Daphne Koller).

As with the previous course, it will be possible to submit assignments in Octave, which offers an opportunity for exposure amongst a wide userbase.

I took the previous course, and I note that Jordi did too. His comments in the forum were very helpful. My expertise in the use of Octave improved enormously.

For those interested, it is possible to register for the course here:

http://www.pgm-class.org/

I plan to take the course and I hope to see Jordi there again!

The instructions for the course specify Octave >= 3.4.0
As you will know, Ubuntu ships 3.2.3

As the course is not mission critical to my career progress, I was planning to use the 3.2.3 shipped with Ubuntu to see if it leads to any problems. Should any arise, I will boot to Windows (shudder) and use the more recent version of Octave from there for the duration of the course.

I was just wondering whether any progress had been made with Debian packages for 3.4
I realise this is undertaken by volunteers and I understand (in principle, not in detail) that there are organisational difficulties for creating Debian packages for Ubuntu, partly (I think; of course I might have misunderstood) through lack of support by them.

Given the opportunity for publicity for Octave amongst a wide audience presented by this course (with the risk of negative publicity if the use of 3.2.3 causes problems), and given that Ubuntu and its various derivatives are currently the most popular Linux distributions (at least according to DistroWatch) I thought that I might mention it here.

I must emphasise that I do not expect any particular difficulty for myself (I mean when finding a suitable form of Octave to use - this is not a claim that I will ace the course!) and I am already a rock solid convert to Octave. :-)

Thanks for all the hard work.
However good you think Octave is, it's much, much better.
Reply | Threaded
Open this post in threaded view
|

Re: Stanford University: Probabilistic Graphical Models

Jordi Gutiérrez Hermoso-2
On 13 March 2012 19:05, pathematica <[hidden email]> wrote:
> For those interested, it is possible to register for the course here:
>
> http://www.pgm-class.org/
>
> I plan to take the course and I hope to see Jordi there again!

I'm quite flattered to receive a personal invitation, and I was about
to accept it, but I came across this:

    https://authentication.coursera.org/auth/auth/normal/tos.php

which appears to be required for signing up to it. This is new. The
machine learning course did not have such extensive legalese.

I cannot in good conscience agree to this. It is asking me to agree to
not copy the course materials, but this is something that would be
very beneficial to other students, for example, setting up a torrent
of the course materials for sharing once the course is done. It is
asking me to not share my code, and makes no provisions for deciding
when sharing my code is appropriate. It says I can't make derivative
works, such as notes from the course that may also help my fellow
students. It repeatedly refers to the annoying meaningless term
"intellectual property".

If they want to make free education, they should make it free. This is
bait-and-switch. "Free," but not free to copy it. "Free" but not free
to share it. "Free" but not free to modify it.

    http://jordi.platinum.linux.pl/piccies/unacceptable.jpg

I urge you to not sign up for the course unless they change their
provisions. Who ever heard of a professor telling you to not create
derivative copies of their work, such as what your course notes might
reasonably be? Who let the lawyers into the classroom?

> I was just wondering whether any progress had been made with Debian packages
> for 3.4

Yes, the Debian Octave Group is furiously at work:

    http://wiki.debian.org/Teams/DebianOctaveGroup

The 3.6.1 release has already hit Debian unstable, so it will probably
be part of the next Ubuntu release.

    http://packages.debian.org/sid/octave

I only hope Ubuntu manages to package the Octave version that actually
works, not the one that is currently in flux.

In the meantime, you can probably get the 3.6.1 low-quality package
that Juan Pablo Carbajal made:

    http://www.octave.org/wiki/index.php?title=Octave_for_GNU_Linux:_Binary_Octave_packages_for_GNU_Linux#Unofficial_binaries

These might not be great, but they should be sufficient for your
purposes. If not, building Octave yourself on Ubuntu is relatively
easy.

HTH,
- Jordi G. H.
_______________________________________________
Help-octave mailing list
[hidden email]
https://mailman.cae.wisc.edu/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Stanford University: Probabilistic Graphical Models

Macy
In reply to this post by pathematica
Thank you for sharing. As a Stanford lifetime alumnus, I will alert them to how inappropriate their approach is.  

Words mean something. And, a person of their word is a person of their word.

One does not protect intellectual property by demanding people adhere to a strict standard of code and then assume these same people will not strictly adhere to that overly strict standard, because it is overly strict!, thus somehow establishing a 'desired' balance to the amount of protection they originally sought! Seems like another example of 'relative ethics' becoming accepted as the norm.

You mentioned lawyers getting into the classroom. Only lawyers would have the brass to say all in the same sentence, "The cost of a free meeting with an attorney is $25." Go to the Santa Clara County Bar of California [where Stanford is located] and seek a free meeting with an attorney. You will find the cost of the 'free' meeting is $25, unless it's gone up.

Regards,
Robert

--- [hidden email] wrote:

From: Jordi Gutiérrez Hermoso <[hidden email]>
To: pathematica <[hidden email]>
Cc: [hidden email]
Subject: Re: Stanford University: Probabilistic Graphical Models
Date: Tue, 13 Mar 2012 21:20:15 -0400

On 13 March 2012 19:05, pathematica <[hidden email]> wrote:
> For those interested, it is possible to register for the course here:
>
> http://www.pgm-class.org/
>
> I plan to take the course and I hope to see Jordi there again!

I'm quite flattered to receive a personal invitation, and I was about
to accept it, but I came across this:

    https://authentication.coursera.org/auth/auth/normal/tos.php

which appears to be required for signing up to it. This is new. The
machine learning course did not have such extensive legalese.

I cannot in good conscience agree to this. It is asking me to agree to
not copy the course materials, but this is something that would be
very beneficial to other students, for example, setting up a torrent
of the course materials for sharing once the course is done. It is
asking me to not share my code, and makes no provisions for deciding
when sharing my code is appropriate. It says I can't make derivative
works, such as notes from the course that may also help my fellow
students. It repeatedly refers to the annoying meaningless term
"intellectual property".

If they want to make free education, they should make it free. This is
bait-and-switch. "Free," but not free to copy it. "Free" but not free
to share it. "Free" but not free to modify it.

    http://jordi.platinum.linux.pl/piccies/unacceptable.jpg

I urge you to not sign up for the course unless they change their
provisions. Who ever heard of a professor telling you to not create
derivative copies of their work, such as what your course notes might
reasonably be? Who let the lawyers into the classroom?

> I was just wondering whether any progress had been made with Debian packages
> for 3.4

Yes, the Debian Octave Group is furiously at work:

    http://wiki.debian.org/Teams/DebianOctaveGroup

The 3.6.1 release has already hit Debian unstable, so it will probably
be part of the next Ubuntu release.

    http://packages.debian.org/sid/octave

I only hope Ubuntu manages to package the Octave version that actually
works, not the one that is currently in flux.

In the meantime, you can probably get the 3.6.1 low-quality package
that Juan Pablo Carbajal made:

    http://www.octave.org/wiki/index.php?title=Octave_for_GNU_Linux:_Binary_Octave_packages_for_GNU_Linux#Unofficial_binaries

These might not be great, but they should be sufficient for your
purposes. If not, building Octave yourself on Ubuntu is relatively
easy.

HTH,
- Jordi G. H.
_______________________________________________
Help-octave mailing list
[hidden email]
https://mailman.cae.wisc.edu/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Stanford University: Probabilistic Graphical Models

pathematica
Thank you for the information about Debian packages and the encouragement to try a local build.

Of more interest are the comments about the terms and conditions of the course. I have to admit, the subtleties of the language were lost on me. I presume that those of you who are active in writing open source are more sensitive to the hidden implications and better at spotting their unacceptable aspects.

Difficulty understanding EULAs is, of course, one of the reasons for using open source and it is yet another reason why those of us who do so without fully understanding the legal implications should be grateful to those that do, and who write the software.

Thanks.
However good you think Octave is, it's much, much better.
Reply | Threaded
Open this post in threaded view
|

Re: Stanford University: Probabilistic Graphical Models

kamaraju
In reply to this post by Jordi Gutiérrez Hermoso-2
Jordi Gutiérrez Hermoso wrote:

> Yes, the Debian Octave Group is furiously at work:
>
>     http://wiki.debian.org/Teams/DebianOctaveGroup
>

Just want to add that the transition to octave3.6 is being tracked at
http://release.debian.org/transitions/html/octave.html

--
Kamaraju S Kusumanchi
http://malayamaarutham.blogspot.com/

_______________________________________________
Help-octave mailing list
[hidden email]
https://mailman.cae.wisc.edu/listinfo/help-octave