Keynote Lecture at NCI by Dr Jeff Ullman

Yesterday I attended a lecture entitled "Algorithm Design for Map-Reduce" by Dr Jeffrey Ullman, Professor Emeritus at Stanford University held in NCI. There was a very good attendance and Dr Ullman discussed (among other things) how mappers, reducers, joins, inputs, outputs, computation costs, communications costs, algorithms, and matrix multiplication work (covered in 62 slides in 60 minutes). All very important stuff I'm sure you'll agree.

Dr Jeffrey Ullman.
Image Source: www.ncirl.ie.

As I had no idea what "map-reduce" was before I went to the lecture I should have anticipated that there was a possibility that I might not know what the subject matter was all about and struggle to understand it. However, I had hoped to find out about this as I had been told that Google use this for PageRanking in searches and that I might get an insight into how it works. 

However, I am slightly embarrassed to report that I did not understand a word that Dr Ullman said during his lecture. I should have walked straight to the corner and put on a Dunce Hat.

I stopped taking notes after about 15 minutes as I had no clue as to what I was writing down. I'm sure that the content of the lecture meant a lot to many of the audience, but without actually explaining what Map-Reduce was I was lost within seconds of the start. Clearly this lecture was meant for a more specialist and knowledgeable audience than me. I am even more embarrassed to say that I Googled "map-reduce" after the lecture and found my true level with MapReduce for Dummies, after which I understood it a little bit better.

Today I attended part of the official opening of NCI's new Cloud Competency Centre. Dr Ullman delivered the keynote speech and this time was most interesting in discussing MOOCs and on-line education. Unfortunately I had to leave after about 15 minutes to go to a class, but the lecture was much more interesting to me than yesterday's one.