Author: Hjalmar Gislason

About Hjalmar Gislason

Founder and CEO of GRID (https://grid.is/). Curious about data, technology, media, the universe and everything. Founder of 5 software companies.

AI Actors in Lord of the Rings

Artificial intelligence, english

As “The Return of the King” is taking the world by storm I think it is suiting to point to this entertaining and informative article from Popular Science last November (after the release of “The Two Towers“). The article explains the AI system that is used to control the characters in the massive battle scenes.

Each orc in a battle has a “mind of his own” making decisions and responding to unexpected happenings. The same goes for every elf, man, uruk-hai, ghost, troll, ent and what have you. The decision trees that are used in the AI system are quite complex, and even though “intelligence” is hardly a word one would directly associate with an average orc, the AI orcs perhaps turned out to be more clever than intended:

“In another early simulation, Jackson and Regelous watched as several thousand characters fought like hell while, in the background, a small contingent of combatants seemed to think better of it and run away. They weren’t programmed to do this. It just happened.” (page 4)

Exams: We Have Changed, They Haven’t

english, Philosophy

I just had a period of exams. Don’t get me wrong, I think I did ok on most of them so this entry is not to blame “it” on something, but I started thinking about the nature of exams, what they are for and how they are performed.

To my best knowledge, exams today are executed more or less the same way they were 100 years ago. The details may vary, but usually it is something like this: The student has a certain curriculum he or she must study to understand and memorize. The exam is then supposed to reflect a statistical example of this curriculum and the result to reflect approximately how much of the curriculum the student has been able to take in.

I can see how this has been a very good format over a century ago, but the times have changed, while exams haven’t. The reality most of us deal with today includes a lot more information than people had to deal with back then and a lot more tools to handle the vast amount of information than was imaginable at the time. I think today’s exams should be more about how well people are able to use those tools to work on their projects than about memorizing various lists that could e.g. be Googled in seconds.
Continue reading →

Merging Social Networks

english, Ideas

Social Software is the in thing these days. Among the sites mapping the social networks of Internet are Friendster, Tribe.net and LinkedIn. Following their success we have less known services such as Ryze, Everyone’s Connected, MeetUp and tons of others. It even seems they are running out of catchy domains with names like itsnotwhatyouknow popping up!

As for myself I’ve watched the development for a long time as I joined sixdegrees.com in 1997 or 1998. It was a different service then than it is now. Actually sixdegrees original patent is still a piece of the action in the race for the throne of the social software world.

Today I’m registered at several of those, but I only maintain my profile on one of them: LinkedIn. It’s simply too much trouble to maintain more than one and LinkedIn has served me fine in so far. But the whole point of this all is kind of lost with all these separate networks. Could there be a way to make these different networks interoperable, so I would only have to maintain one profile?
Continue reading →

BioBloc: Do-It-Yourself Animals

Artificial life, english

BioBloc is an extremely cool website. The site allows you to build your own three dimensional animals from blocks and then use genetic algorithms to evolve them to learn to walk, jump, run, roll, etc.

The application – made as a java applet – is very easy to use and it has a pretty good help system to get one going.

It reminds me a lot of Karl Sims‘ Evolved Creatures that Wetware wrote about in September, but now we can play ourselves. I urge you to give it a try. It’s tremendous fun.

(Thanks to Joi for the tip)

Interesting Content Indicator

english, Ideas

Every day I read a lot of news, articles and other information online. Most of it I read through a news aggregator (have tried a few, currently evaluating NewzCrawler), but I also visit a few websites regularly (that don’t have RSS feeds) and people send me interesting links via Instant Messaging or email.

The tips I receive from friends and Wetware readers and the links I find on my favorite blog sites tend to be the most interesting material. The reason is simple. These are hand picked links from people that share (or know) my interests. If I receive a link from a friend via IM, I will always read it. I know they wouldn’t “interrupt me” unless it was something that they knew I would find interesting.

With the sea of information out there, we will increasingly need better mechanisms to make sure that the interesting content “floats” to the surface and using some sort of collaborative filtering seems like a good approach. Obviously several such solutions have been attempted, but none of them has really done the job. Here’s my suggestion:
Continue reading →

A Collaborative Approach to the Turing Test

Artificial intelligence, english

Despite that its importance may be debated, the Turing Test at least poses a very hard and interesting computer science problem: How to build a program that can engage in a text conversation with a human being so that the human cannot tell if it is a computer or another human it is talking to?

This is at least the common interpretation of the Turing Test although the imitation game that Turing put forth in “Computing Machinery and Intelligence” actually involved a text conversation where a computer would do as well as a male human in making an interrogator believe that it/he was a woman (which is actually a bit different as it means that both players are imitating – claiming to be something they’re not).

This problem has proven harder to solve than probably even Turing himself realized, and many different solutions have been attempted, most of them failing quite miserably (this is the state of the art). Well, everybody has their own plan on how to get rich – that fails – so here’s my suggestion.
Continue reading →

The Turing Test and Extrasensory Perception

Artificial intelligence, english

Having been interested in Artificial Intelligence for a long time, I’m almost embarrassed to admit that I hadn’t read Alan Turing’s famous article: “Computing Machinery and Intelligence” until today. This is the article where the Imitation Game – later known as the Turing Test – is put forward.

Although I was familiar with most of Turing’s arguments there, reading it was nevertheless truly inspiring (more on that in a moment). Turing’s writing style is brilliant.

In part 6 of the article, Turing explores several contrary views to the notion that machines can think. Many of these are still today hot debates in AI discussions. One of them however, and actually the one that Turing seems to find the strongest one, struck me as quite odd. How could a thinking machine ever account for extrasensory perception such as “telepathy, clairvoyance, precognition and psychokinesis” for which “the statistical evidence, at least for telepathy, is overwhelming”?
Continue reading →