On Saturday the 31st of October, I’m going to be the speaker at Extrobritannia here in London. I went along to their last meeting and it was totally packed out, nearly a hundred people I believe. Having both Dr. Aubrey de Grey and Dr. Anders Sandberg speaking explains why!
I’ll be covering topics from my PhD thesis, such as the definition of intelligence, Solomonoff’s model of Induction, Hutter’s AIXI and more recent work such as the Monte Carlo approximation of AIXI by Veness et. al. I’ll also include a few thoughts on how recent discoveries in theoretical neuroscience might help guide work towards AGI.
You can find the event on facebook here, and the announcement of the event on the Extrobritannia list here.
Dressing as a witch, wizard, ghost, etc. is optional.
Looking forward to this!
Lot of interest from my group so don’t think your not gonna get packed out!
Cya Saturday
Yeah, it should be fun
Intelligence, complexity theory, AIXI, the machine learning algorithms in the brain… I’ve even managed to get a video clip of AIXI in action.
more hype from facebook about you:
Mark Tanner
Unfortunately, instead of pursuing AI for sensible applications, they have insisted on putting MCU’s in toasters, which try to anticipate how brown you would like your toast, but always get it wrong.
We should make the yardstick for AI a test based on an observers ability to distinguish toast made by an ‘intelligent’ toaster and toast prepared by a human being using a grill. [/sarcasm]
Miguel Sanchez
You know what thats not as crazy as you think!
Me and LAurie reckon any machine which can take part in the Japenese Tea Ritual observing all ettiquite would be super AI – Have a look at wikipedia to see how complicated the ettiquit is!
Mark Tanner
I was simply suggesting a ‘bare bones’ approach to the Turing Test..
The problem always seems to be one of logic – Most attempts at AI seem to employ some kind of logic.
No matter how advanced the implementation of artificial logic, it fails to address the problem that very little about human behaviour is actually logical. The Japanese tea ritual… Read more is an excellent example. The goal should not be to get a machine to replicate the ritual – The true breakthrough would be to get that machine to come up with the ritual in the first place.
Miguel Sanchez
oh man – i fkn wish you were coming down saturday – you would love it – will send youtube link when they post it -
ooh pls edit names + swearing shwn – my bad
Sorry i missed you shane; i did read your thesis and some other publications up for it.
I hope to get a report of what you said or watch the video.
I think there are many models of intelligence and the question for me is what does one want to do with it/
See my new post: it has the slides and links to the video on youtube.