Wow this is space age stuff. Combining hard logical inference with probability is very exciting. We can codify background knowledge like physical laws to speed up the process of learning. At the moment NN have to have tons of training data just to pin down various truisms we already know. They learn from a blank slate everytime which is not exploiting all the background knowledge we already know. Well, convolution neural nets kind of encode transnational invariance in their architecture, but we should be able to be more general, this is the path to that!
Nice article. I haven't experimented with PowerLoom in many years. I will check it out again. I have a JRuby wrapper for PowerLoom on github and I should make sure that works with the latest snapshot.
BTW, there was a mention of PowerLoom on the iPad, but I searched for it in the Apple Store and couldn't find it.
Very interesting article, but the page layout at the top of the page seems broken. There's a huge blob of article titles taking up the first page and a half, followed by a big picture of books. Ends up making me scroll almost two pages just to get to the first heading.
Great article, why make a fundemental distinction between isa and association? any realtion can be modelled as isa.
eg isa dog and isa agent who plays frisbee.
Any such relation can easily be modelled in a powerful relational database such as postgres, including recursive or ancestral, so why use this strange niche tool.
