My Photo

Your email address:


Powered by FeedBlitz

June 2009

Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30        
Blog powered by TypePad

« When Risk Assessment is the Risk | Main | Hacking the 2008 UK Ironman: Kick to Kill »

September 28, 2008

The Fast Last Puzzle Piece

The notion that the more data, the slower the system – ain’t always true.


My favorite way to explain this very important phenomenon involves the familiar process of assembling a jigsaw puzzle.


The first piece you take out of the box and place on the work surface requires very little computational effort.  The second and third pieces require almost equally insignificant mental effort.  Then as the number of pieces on the table grows the effort to determine where the next piece goes increases as well.  But there is a tipping point where the effort to determine where to place the next piece gets easier and easier … despite the fact the number of puzzle pieces on the table continues to grow.


Well isn’t it interesting, although obvious, that those last few puzzle pieces take nearly as little effort as the first few!


I have witnessed this.   


This has a slew of ramifications.


This does not apply to all domains.  This behavior requires: (a) observations from the same universe; (b) observations with enough features to enable contextualization; (c) observations in which these features can be extracted, enhanced and classified; (d) sufficient saturation of the observational space; and (e) enough smarts to stitch these puzzle pieces together.


Context accumulating systems, fed appropriate observations, can be expected to have this behavior.

 

RELATED POSTS:

More Data is Better, Proceed With Caution

Context: A Must-Have and Thoughts on Getting Some …

To Know Semantic Reconciliation is to Love Semantic Reconciliation

Big Breakthrough in Performance: Tuning Tips for Incremental Learning Systems

 

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d83452946769e2010534d46a7a970b

Listed below are links to weblogs that reference The Fast Last Puzzle Piece:

Comments

I like using the notion of perimeter. It starts from scratch, increases to a tipping point after which each new piece is decreasing the perimeter rather than increasing it and rapidly decreases. Software tends to be similar IMHO, not just with searching information but also programmer effort in using APIs. Designing software so that it has maximum utility with minimum perimeter is an ideal I aspire to.

This happens because when the memory in the system is full (when it contains the most data in the middle) it will take the longest to process the data. I really like your comparison to a jigsaw puzzle, because that makes tons of sense.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Working...
Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.

Working...

Post a comment