ArabicLexikon Game – ImagineCup 2007
A few days, I received an email from Samer Chidiac of Microsoft stating that my team is qualified for Round 2 of ImagineCup 2007 (Software Design) which will be held this Saturday 12th of May in German Jordanian University in Amman.
We (my partner Khalil and I) originally intended to participate in Web Development category, but unfortunately we missed the deadline so instead we decided to switch to Software Design (obviously we had to change the whole website to a Windows Forms application).The result is pretty good, we named it ArabicLexikon (with a ‘k’ !!).
The concept is pretty simple, ArabicLexikon is a ‘game with a purpose’ that aims to build a lexicon knowledge base for Arabic language. It was inspired by ESP Game which aims to label all images on the web.
I promise to post an extended explanation soon plus the source code once the project is finalized. In the meantime feel free to check out the PPT presentation we prepared for this event. You can read more about the game below or as a PDF File:
ArabicLexikon Game Overview
We introduce a game that is fun and can be used to create useful output.
ArabicLexikon is a ‘game with a purpose’ that aims to build up a lexicon knowledge base for Arabic language. It was inspired by ESP Game (http://www.espgame.org) which aims to label all images on the web.
GENERAL DESCRIPTION OF THE SYSTEM
The game is played by two partners and is meant to be played online by a large number of pairs at once. Partners are randomly assigned from among all the people playing the game. Players are not told who their partners are, nor are they allowed to communicate with their partners.
The game consists of asking the players questions about randomly chosen words and recording their answers. The design of the game ensures that data obtained through the game are correct. The actual implementation is described below.
Questions can be: What is the stem for the word “Game”? What is the classification of this word (verb, noun, etc…)?
USEFULNESS OF THE SYSTEM
Our system makes a significant contribution because of its valuable output and because of the way it addresses problem of gathering the data. Rather than using computer algorithmic techniques, which don’t always work well enough, we encourage people to do the work by taking advantage of their desire to be entertained. By playing the game, people help us collect data not because they feel helpful, but because they have fun. And ultimately the data generated by our game can be useful for a variety of applications such as being used as facts to further testing and strengthening existing algorithms.
This game can be considered a “human algorithm”: given a word as input, it outputs a set of facts related to the word. Instead of using a computer processor, this “algorithm” uses ordinary humans interacting with computers throughout the Web. Our system therefore significantly contributes to HCI in two ways: it collects data that can improve HCI applications, and it addresses a typical AI problem with new HCI tools.
HOW DOES IT WORK?
- Players sign in to the game
- The engine randomly pairs two players
- The game randomly select a word for the two paired players
- The game asks the players a certain question about the word, and allow them to enter their guesses
- The game records the players guesses until a match is detected or a timeout is reached
VARIATIONS OF THIS GAME
This game can be implemented two ways:
- In a symmetrical way were the two players are given the same word, and when the players enter the same answer a match happens
- In other cases, it can be implemented in an asymmetrical way where one player’s output is the other player input and he must guess what word the first player was given (given some hints like starting letter or number of letters).
EXAMPLES OF POSSIBLE GAMES
- Words classification game
- Words relations
- Words stem game
Attachments: PDF, PPT