... or maybe not. I have always thought that when some scientific discovery looks too fantastic to be true, it usually is. Even more so if it's published in the press.
This does not seem to be the case for robot scientist
, and also found National Geographic, no less
), even as the article mentions stuff such as "advanced artificial intelligence" and "sophisticated robotics" (meaning usually that the journalist can't make head or tails of it).
Feeling my job at peril, I have dug down a bit into the news. The mentioned scientist home page doesn't help
, or even his publications page
, since it's not been updated since 2001. Checking ScienceDirect
only yields a 3-year-old paper.
However, the ailing ResearchIndex
does help. I find two papers: this one, which is very old
(and I maybe could have found it somewhere else, but at least here, I can follow citations), and ILP: Just do it
, a review on Inductive Logic Programming. There's also an EC-financed network of excellence
dealing with the subject. Guess it wasn't the robot who wrote the grant application.
I might be off mark here, but my guess is that what they have done actually is teach some inductive logic programming module, which seems to be just like the old expert systems, but with probabilities added, stuff on bioinformatics, and then connect it to some bio thing, maybe a sequencer or somesuch.
So, there's still room for the old kind of scientist.