4 ms·
I think what's changed since this was posted in 2011 is the emergence of embeddings and the need to take advantage of its higher dimensional space. While embedd
by jdonaldson 1y ago
I think what's changed since this was posted in 2011 is the emergence of embeddings and the need to take advantage of its higher dimensional space. While embeddings expose more underlying structure that can be used for tensor math, ranking systems often are still good ol' trees. This project to me points at a new major "hinge" of information architecture.
- bawolff 1y agoWhat does this have to do with wavelet trees? (Sorry for being a dick if im wrong) - was this an AI generated comment that got confused by the domain specific meaning of the word "rank" in this context?
- quantadev 1y agoYou weren't wrong. Wavelet Trees have no relationship whatsoever to vector embeddings.
- jdonaldson 1y agoThat is a pretty badly overloaded word for it, and I didn't even pay much attention to the notion of "rank" anyways. I'm mainly interested in how the text is represented with bit vectors. It's very reminiscent of how vector math plays out in other ML domains, but I would bet that many people working with text have never heard of it.