Fortnightly Mailing

Seventh report from the Norvig/Thrun/Stanford/Know Labs Artificial Intelligence course

(Other posts tagged ai-course.)

Substantial  changes made to A3 and small changes made to A5 and A7 below, 21 November 2011, and to introductory paragraph, 23 November 2011. [For visitors from the Aiqus discussion about student numbers, note that more light is shed on the "YouTube video counts question" - a distracting side-issue - in paragraph seven of report number two, 18 October 2011.]

Here is my seventh participant's report from the Stanford Introduction to Artificial Intelligence course. It is in three parts. A is the report proper. B picks up on a pre-arranged call from Sebastian Thrun. C lists further free CS courses that will be available from Stanford University in January 2012.

A - Report

1. A lighter week from the point of view of studying, because the last week saw less material presented, to allow students time to prepare for the midterm "examination", aka the midterm.

Using AI in formative and summative assessment

[Decided to remove image of scratch pad shortly after publishing this.]

In September 2007 I had a personally revelatory moment concerning how AI (as I now know it to be) might be used to provide formative feedback for learners. It stemmed from being involved in running the 2007 ALT conference at which Dylan Wiliam and Peter Norvig each gave keynote speeches. With Richard Noss (who now directs the ESRC/EPSRC funded TEL programme) we set up a Google Doc "scratch pad" to gather a shortlist of issues that could be worth examining. Not a lot (well, nothing!) came from it, as is often the way.

Doing the Stanford AI course (see also my piece in ALT News Online) has sensitised me to current work on the same thing.

Here are some examples. If you know of others, please post them as comments and I will collate them.

• How Khan Academy is using machine learning to assess student mastery, November 2011 piece by David Hu.
• The elements of e-assessment, 2011 article by Sophie White of the examining body OCR (Oxford, Cambridge and RSA Examinations).
• Edexcel chief puts hi-tech agenda to the test, 2010 TES article about developments in Pearson's examining body Edexcel.
• Massive scale data mining for education, November 2010 piece by Greg Linden in Communications of the ACM (via PN).
• Computer analysis of test-takers' answers to standarised tests is revealing cheats, May 2011 article in the Economist mentioning the "relentless logic of combinatorial statistics". [Thanks to DM for the link.]

Fifth report from the Norvig/Thrun/Stanford/Know Labs Artificial Intelligence course

The paths of blindfolded walkers trying to walk in a straight line in overcast (blue) and sunlit (yellow) conditions. From Unit 8.2 of Introduction to Artificial Intelligence course.

 

(Other posts tagged ai-course.)

Here is my fifth participant's report from the Stanford Introduction to Artificial Intelligence course.

1. I'd finished the work before I noticed that, unannounced, the order of the course had been changed from that shown in the originally published outline.  Thus units on Representation with Logic, and Planning, taught by Peter Norvig, have come before the originally scheduled units on Hidden Markov Models and Bayes Filters.

2. This change of order probably explains why this week's study felt somewhat disconnected from last week's,  a fact emphasised by a change of teacher from Sebastian Thrun to Peter Norvig. Thrun, it has to be said has a less austere and more down-to-earth presence than Norvig, whose delivery style is dry and very concentrated. Underlying this, I think Norvig's material on propositional logic and on mathematical representations of plans is by its nature relatively abstract: and for me this spells trouble, being someone who has always tended to struggle with the abstract. 

3. The advice I'd give from a course design point of view is to further strengthen the illustrations as to why these kinds of abstraction matter.  The video from which the picture above is taken a good example of this:

4. Secondly this kind of more abstract content needs more rather than less use of "dialogic" check questions, as has been the case in other units of the course. To illustrate this point here are the contents-lists of Units 4 and Units 8. Check-questions are indicated by ?. 12 sets of check questions out of 21 sections is a much more promising ratio than 3 out of 22.

5. During the last week Sebastian Thrun and Peter Norvig ran an online "office hours" session using a Google Plus "Hangout". I was not around to try to join this, but from Thrun's "We apologize for the large number of people who were denied participation in the online office hours via youtube. We had a lively discussion which was recorded on video - mostly on topics beyond this course (e.g., what are great research topics). We will soon post the video on this site. Apologies again. Technical problems with the Hangout-Youtube link." it looks as if this was a partial success. I think this is a "forgivable" issue, given the very large number of people who will have attempted to take part in it.  

6. Next week I hope to find out whether and if yes by how much the participation rate has changed, as measured by the submission of homework for weeks 3 and 4.

Fourth report from the Norvig/Thrun/Stanford/Know Labs Artificial Intelligence course

(Other posts tagged ai-course.)

Small edits made 30/10/2011,  some of which are indicated.

Here is my fourth participant's report from the Stanford Introduction to Artificial Intelligence course.

This is a shorter report than #1-3, mainly because the course has got into a rhythm and because there've been no substantial changes in delivery methods.

1. Despite the work that has been done to the web systems that sit behind the site, it looks as if there were again overload issues at and around the week 2 homework submission deadline, and this despite the probability the number submitting homework may have dropped quite a bit dropping by nearly 20% to ~37,000 from the ~46,000 reported after the week 1 homework deadline.

2. The course continues to fascinate. For example it is nice to gain a practical understanding of how things like spam filters actually work.

Third report from the Norvig/Thrun/Stanford/Know Labs Artificial Intelligence course

(Other posts tagged ai-course.)

Here is my third participant's report from the Stanford Introduction to Artificial Intelligence course.

For more than half my working life I have been engrossed by on-line distance learning (yes - since 1992). I've been unable to resist giving the course organisers a piece of my mind about aspects of the underlying course design, having been given active encouragement to do so. I've pasted at A below an excerpt from some feedback provided to Know It!'s David Stavens earlier this week.

The rest of this report concerns my experience of the course as a learner over the last week.

Second report from the Norvig/Thrun/Stanford/Know Labs Artificial Intelligence course

Screenshot from AI Course Unit 2, Topic 34, Note on Implementation

(Other posts tagged ai-course.)

Final paragraph updated 18/10/2011, and 23/11/2011

Here is my second participant's report from the Stanford Introduction to Artificial Intelligence course.

1. Over the last week I worked through the second section of the course - Problem Solving - consisting of nearly 40 short, low-tech videos. I described the pleasing, quirky, design of these in the first report.

2. Here you can review the Problem Solving materials, without being logged into the course. If you click on "cc" in the video-control bar you will see how the videos have been captioned and translated in multiple languages (by volunteers, using the dotSUB platform).

Why give a course away online for free?

Rhodri Marsden asks in the 12/9/2011 Independent why Stanford University is "giving away a course online at no cost".

Marsden's piece is wide-ranging, and it draws on the knowhow of several people in UK HE, including Nial Sclater and Susanah Quinsee.

But it was filed some time ago¹: if Rhodri had done a bit more digging at the time, the possibility that Stanford or an associated business might be looking to pilot a delivery method for future commercial or revenue-generating use might not have escaped him.

Some quick observations:

RIP Michael S Hart

Image source

9 September 2011

RIP Michael S Hart (who founded http://www.gutenberg.org/, in effect inventing eBooks). Gregory Newby's obituary will make you think, and hopefully donate to Project Gutenberg.

[Added 16 September 2011]

You may also want to read Martin Childs's full and fullsome obituary in the Independent.

Many pieces loosely joined - the Scottish Government's interesting take on technology in learning

This short talk by Michael Russell MSP - published on 8 September - is definitely not a run-of-the-mill perspective on technology enhanced learning:

The Norvig/Thrun/Stanford free online AI course: more transparency needed concerning the relationship between Stanford University and KnowLabs

(Other posts tagged ai-course.)

Peter E. Murray pulls together several related comments and questions about the relationship between Stanford University and KnowLabs, the "angel-funded" start-up that is closely associated with the free online course that has attracted mass-attention (and mass enrolments) in recent weeks.