Skip to main content

I watch movies, do puzzles, love my family and try to keep up with science and technology

Google TensorFlow and IBM Watson

1 min read

How do they compare what can you do with either of them ?

The #MOOC: Question Answering Technologies Behind IBM #Watson

1 min read

The last couple of weeks I have been following this MOOC

IBM Watson Academy

And today I completed the course


This MOOC gives a good understanding of what it takes to build Watson and apply it to different domains.

From the abstraction level it lies between:

Cal Tech's learning from data course:


IBM's training for Watson Explorer Analytical Components.

The good news is, the UIMA pipeline training can (and will) pay off when you start working with Watson as a cognitive system.

The other news is that language support is still a challenge for smaller language, but than again Analytical Components supports much more languages than Watson Cloud does.

When you want to get inspired, just watch the videos in the epilogue (lecture 9).


While we wait for the official support for Dutch, we will start learning UIMA by using Analytical Components and applying it to dutch text.

Aflevering de volmaakte mens #dvm geeft een mooi perspectief op #MachineLearning en #AI

1 min read

Geinteresseerd in en

De aflevering van 10 juni 2015 waarschuwt voor mogelijke scenario's maar is niet alleen pessimistisch, zie hier

The AI control problem and the tales of the Genie in the lamp

1 min read

Inspired by this weeks Triangulation (177) episode about Super Intelligence, it occurred to me that the control problem has many similarities with the 3 wishes problem that come with the Genie in the lamp.

Those tales are full of warnings and almost all the time the wishes don't worked out as intended.

The smartest first wish for the Genie would be:

Please fulfill my next wishes the way I intended them to be...

Since Genie is so much smarter he/she should be able to understand this.

Based on the triangulation show with nickbostrom I am inclined to think that the AI control issue should be addressed in the same way.