Review: edX Microsoft: DAT263x Introduction to Artificial Intelligence (AI)

As i have said before: I have had an interest in Artificial Intelligence for a long time, and over the years I have used bits of it in projects, but never really studied it.

At the moment I and some others are toying with project  ideas that use Artificial Intelligence. I borrowed a book “Introduction to Artificial Intelligence” by Ertel  to brush up on my understanding

Tharindu pointed to an edX course Introduction to Artificial Intelligence and I decided to give that a go.

I hadn’t taken an edX course before so I had to go through the rigmarole of setting up an account.

The course was divided into four topics:

  • Machine Learning
  • Language and Communication
  • Computer Vision
  • Conversation as a Platform

With the suggestion of  studying one per week and an effort of 3-4 hours per week.

Machine Learning was divided into:

  • Regression
  • Classification
  • Clustering

For this topic I watched each of the Fundamentals videos (at 1.25 speed as that is clear for me) and I followed the 26 page lab. This took about 5 hours elapsed time, and I feel it would be difficult to do it any quicker. The videos were straightforward as I knew of the techniques, the lab was very much a step by step guide, and I only once went wrong and quickly found what I had done mis-copied. I have now realised there were a set of videos on how to use the Azure Machine Learning Studio, I have only watched the first of these – not sure how I missed the others – watching them would have taken another 50 minutes or so and I doubt it would have helped me do the lab quicker.

Language and Communication was divided into:

  • Getting started with text processing
  • Introduction to Natural language Processing
  • Language Understanding Intelligent Services

For this topic I watched all the videos and attempted the lab .

Again I already knew something about the topics so I guess it was easier for me but the whole thing took longer than suggested. I was finding the lab straight forward until I got an error message:

Error: nonnumeric port: ‘//westeurope.api.cognitive.microsoft.com/text/analytics/v2.0’

After copying a url from one browser window to another, my debugging skills could not help me get this problem solved, and so I had to abandon the lab – I tried the discussion boards and Google but no clue as to what is astray. With more time I would restart and use a different browser, but I think for now the lab has served its purpose.

Having written the above I decided to revisit and determined that I was only suppose to use the base ‘westeurope.api.cognitive.microsoft.com’ not the whole url the interface offered to copy for me.

Computer Vision demonstrated lots of things that the apps available from Microsoft can do. It is amazing how this field has advanced since I first knew of work in this area. I watched all the videos and did the lab, I was a little cheesed off with the review questions as two of the three asked:

“Which Microsoft Cognitive Services API should you use?”

which seemed closer to understanding their name choices than AI (maybe that is because I got both those questions wrong having got everything else right!).

Conversation as a Platform was an interesting topic, introducing the concept of Bots (Chatbots) and Intelligent Bots. The lab followed the videos and were designed to enable learners to make and deploy bots.  I was enjoying this and thinking how straightforward it was to use until I hit a problem. The step included instruction to set and App ID and password automatically:

  • Microsoft App ID and password: Auto create App ID and password.

When I attempted to create the App I kept getting back:

  • Authorization_RequestDenied

Which the help suggests is because I have insufficient privileges, and that I could set one up manually – but that route also says I have insufficient privileges – since  I have no idea how to make my Free Trial account more privileged so I am stuck here! I have posted on the course discussion board a query re this – but the discussion board seems to have very few posts so I am not optimistic I will get a response.

Having written the above I thought I should retry the steps in the hope that it would work – but it didn’t – I am still getting the same error message.

So moving on: the review question were again very Microsoft centric – asking which service to choose – not testing my understanding of AI – but I had paid attention and this time got them right!

The course finished off with a what to do next.

I enjoyed the course and it gave me a taste of modern AI, I enjoyed the use of the Microsoft tools, the final authorisation problem will stop me with some experiments I have in mind to try.

Would I recommend the course to a friend? Well it would very much depend on who that friend was and what their background is.

Book #32 Introduction to Artificial Intelligence by Wolfgang Ertel

I have had an interest in Artificial Intelligence for a long time, and over the years I have used bits of it in projects, but never really studied it.  At the moment, along with others, I am toying with project  ideas that use Artificial Intelligence. I borrowed this book from Richard to brush up my understanding.

This book is too much of a text book (aimed at Computer Science undergraduates) to be very readable, but skimming through helped remind me of topics.  The book was originally written in German, and the translation to English left examples with German roots. The individual chapters can be approached with limited dependence on earlier material.

The topics covered include: Propositional Logic, Predicate Logic, PROLOG, Problem Solving, Uncertainty, Machine Learning, Data Mining, Neural Networks.

I think this book is VERY niche, it is unsuitable for anyone who wants a general introduction to AI. It is ok as a text book, it is has a lot of details that would be covered in an undergraduate course, and would be good to supplement  lectures.