John Rauser: How humans see Data

Excellent talk on data visualization. The abstract from YouTube:

John Rauser explains a few of the most important results from research into the functioning of the human visual system and the question of how humans decode information presented in graphical form. By understanding and applying this research when designing statistical graphics, you can simplify difficult analytical tasks as much as possible.

Selected links:

Text segmentation and coding for QDA

In my second master thesis I was trying to automate part of the process required for public policy design. The results were not very satisfactory in terms of the applicability in practice. I hope to someday return to the idea and utilize machine learning methods which became available since then (hm, deep learning maybe) + knowledge and experience gained during doctoral studies at CLEA and ECCO.

Veitas, V. Text Segmentation and Coding for Qualitative Data Analysis. Master Thesis. Katholieke Universiteit Leuven. 2009.


Joscha Bach on How To Build A Mind

AI researcher Joscha Bach gave a beautiful presentation at 30C3 (Chaos Communication Congress) “How to Build A Mind” in December 2013. It seems that the talk was somewhat aimed at getting hackers interested in Artificial Intelligence project.

Here are the most provoking /  interesting / informative thoughts of the presentation (well, according to my interpretations, which are in the brackets []) plus approximate times of the video where you can find/check them:

09:11: There is no need to fear for the AI. It is already there in the form of corporations (kind of artificial legal persons). The “real” AI will not be worse for sure.

?: God is a system architect, devil is a hacker ;);

19:00: Intuition is overrated. In philosophy, intuition matters, which usually means that it matters what audience think is right. In computer science what matters is whether your software run. [I guess the idea is that software is initially based on intuition of the architect, but nobody cares about intuitions when it does not run..]

24:14: Computational models of the brain:

  • Compartment models [or multi-compartment models, which are somewhat similar to the well-stirred reactor vessels of the Chemical Organization Theory);
  • Integrate-and-fire model (some sort of simplification of the former);
  • Artificial neural networks (very crude simplification of the former !).

34:26: A short history of AI. There was a comment around Wiener and Ashby that “second-order cybernetics with Maturana and stuff” which turned to humanity was an unfortunate development…

Screenshot from 2014-01-03 22:53:57

?: Artificial Intelligence project is about using Computer Science to understand a mind (said Marvin Minsky sometime ago when coining the term “AI”).

37:03: Short history version #2.

Screenshot from 2014-01-03 23:01:03

47:36: Affective states (feelings, etc.) are configuration of the cognitive system.

49:11: Simplified list of problems

Screenshot from 2014-01-03 23:05:47


49:46: What you need to learn if you want to gen involved into AI project:

Screenshot from 2014-01-03 23:11:47


51:10: Details of representation problem

Screenshot from 2014-01-03 23:07:43

52:06: Details of perception problem

Screenshot from 2014-01-03 23:09:24

54:08: Integration of everything [something similar to what Ben Goertzel calls Cognitive Synergy, I guess].

Screenshot from 2014-01-03 23:15:19