0

I'm reading a paper and in a part of this paper there is a note about intersecting sets using address bus. This is the exact quote from the paper:

Fast retrieval methods often rely on intersecting sets of documents that contain a particular word or feature. Semantic hashing is no exception. Each of the binary values in the code assigned to a document represents a set containing about half the entire document collection. Intersecting such sets would be slow if they were represented by explicit lists, but all computers come with a special piece of hardware – the address bus – that can intersect sets in a single machine instruction. Semantic hashing is simply a way of mapping the set intersections required for document retrieval directly onto the available hardware.

I have some basic knowledge about cpu architecture. All I need is an abstract explanation to understand how this operation is done.

P.S. The paper is about the sets, but my question is general (any kind of data).

Matt
  • 22,721
  • 17
  • 71
  • 112
Pouya
  • 1,266
  • 3
  • 18
  • 44
  • We're missing the context. It says "such sets", but we have no idea what it's referring to. Without that, there's no way to tell what this means. (I found the paper, and I still can't tell what they mean.) – David Schwartz May 18 '12 at 10:31
  • @DavidSchwartz As I've written in the post script, my question is general and not data-related. But in any case, I edit the post. What you see in the quotations now, is the complete subsection of the paper. – Pouya May 18 '12 at 10:37

1 Answers1

0

Essentially, what he's saying is that you can implement any mapping of input numbers to output numbers in a single instruction if you have enough memory. Simply populate memory with your mapping and from the read the address in your mapping corresponding to the input number.

David Schwartz
  • 179,497
  • 17
  • 214
  • 278