Re: scoring differences between bitmasks

From: Ben <bench(at)silentmedia(dot)com>
To: Martijn van Oosterhout <kleptog(at)svana(dot)org>
Cc: Postgresql-General mailing list <pgsql-general(at)postgresql(dot)org>
Subject: Re: scoring differences between bitmasks
Date: 2005-10-02 19:53:53
Message-ID: 26BBC302-2BD1-4068-B7AE-8E42D1FC72C0@silentmedia.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Hrm, I'll think about it. Thanks!

On Oct 2, 2005, at 12:50 PM, Martijn van Oosterhout wrote:

> On Sun, Oct 02, 2005 at 10:17:10AM -0700, Ben wrote:
>
>> Yeah, I thought about this, but damn my 3-dimensional mind, I just
>> can't convince myself this will work for 256 dimensions. :) Or
>> rather, I can see how it could, *given* good placements of the grid
>> points to snap each vector to. Unfortunately, I don't know enough
>> theory to know what good placements would be.
>>
>
> My intuition tells me that just like real life, you want them to be as
> far apart as possible. That's why I picked those strings. If the
> string
> is length N, all those key strings are distance n/2 apart from
> eachother.
>
> As for proof, I think it will work due to the triangle inequality
> holding true. Try not to worry about the dimensions, if you don't
> assume the number of dimensions, it'll work for any.
>
> My Google search hasn't come up with anything, although searching for
> "binary vector space search" comes with some almost but not quite
> results.
>
> Good luck...
> --
> Martijn van Oosterhout <kleptog(at)svana(dot)org> http://svana.org/
> kleptog/
>
>> Patent. n. Genius is 5% inspiration and 95% perspiration. A patent
>> is a
>> tool for doing 5% of the work and then sitting around waiting for
>> someone
>> else to do the other 95% so you can sue them.
>

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Magnus Hagander 2005-10-02 20:31:46 Re: Portable PostgreSQL
Previous Message Martijn van Oosterhout 2005-10-02 19:50:42 Re: scoring differences between bitmasks