Geek Answers Handbook

What optimization problems do you want to solve?

I like working on AI optimization software (genetic algorithms, swarm of particles, Ant colony, ...). Unfortunately, I was not able to solve interesting problems. What problem would you like to solve?

+7
artificial-intelligence mathematical-optimization
source share
7 answers

This list of NP-related issues should keep you busy for a while ...

+12
source share

How about a Hutter Award ?

From a Wikipedia entry:

The Hatter Prize is a cash prize funded by Marcus Hatter that rewards data compression improvements for a specific 100 MB English text file. [...]

The Hatter Prize's goal is to encourage artificial intelligence (AI) research. Organizers believe that text compression and AI are equivalent tasks.

Basically, the idea is that to create a compressor that is capable of compressing data more efficiently, the compressor must, in the words of Marcus Hatter, be smarter. For more information on the relationship between artificial intelligence and compression, see the Motivation and FAQ sections of the Hutter Prize website.

+7
source share

Does Netflix Prize Count

+4
source share

I would like my bank balance to be optimized so that as much money as possible is left at the end of the month, and not vice versa.

+3
source share

How about a go game ?

+2
source share

Here is an interesting practical problem that I encountered, tinkering with color quantization and image compression.

The basic idea is that I need a program to which I give a picture, and it reduces the number of colors as much as possible without my comment. Since each person has a different sensitivity of the eye (and the eyes have different sensitivity of the intensity of red / green / blue), it should be possible to somehow determine this threshold of sensitivity.

In other words, in the truecolor image, replace each pixel color with a different color so that:

  • The total number of different colors in the picture would be the smallest; and
  • Each new pixel will have a color no further from the original color than the user-defined value D.

D can be defined in different ways, choose your favorite. For example:

  • Separate red, green and blue components to determine the maximum possible deviation for each of them (for each pixel you get a rectangular cuboid of permissible replacement values);
  • A real number that will represent the maximum allowable distance in the RGB cube (for each pixel you get the scope of the allowable replacement values);
  • Something between them or something completely different.
+2
source share

The most effective solution for a given set of Sudoku puzzles. (excluding brute force methods)

0
source share

More articles:


All Articles