/* ---- Google Analytics Code Below */

Saturday, August 25, 2018

Looking at Sorting in Multiple Spaces in a New Way

I did much work in the area of sort algorithms, so this is interesting. Its not quite conveyed why this is useful.  We do know how to sort very disparate data very quickly and every time you press a smartphone key you initiate a number of fast sorts.   What is new and remarkable is how sort algorithms can interact with broader domains and dimensions of information. 

Not very technical,  but technical papers on the concept linked to below.

In QuantaMagazine:

Universal Method to Sort Complex Information Found

The nearest neighbor problem asks where a new point fits in to an existing data set. A few researchers set out to prove that there was no universal way to solve it. Instead, they found such a way.

By Kevin Hartnett,   Senior Writer Quanta

" ... Now, a team of computer scientists has come up with a radically new way of solving nearest neighbor problems. In a pair of papers, five computer scientists have elaborated the first general-purpose method of solving nearest neighbor questions for complex data.

“This is the first result that captures a rich collection of spaces using a single algorithmic technique,” said Piotr Indyk, a computer scientist at the Massachusetts Institute of Technology and influential figure in the development of nearest neighbor search. ... "

Technica papers background:
https://www.ilyaraz.org/static/papers/spectral_gap.pdf
https://ilyaraz.org/static/papers/daher.pdf

No comments: