|LMC2 General Information||
|LMC2 consists of a tournament, in which all the submitted programs will play against a judge/trainer in a number of different language games. The games won't be revealed to you, so you will need to come up with a program which can learn to play ("converse") an unknown game ("language").
|How does it work?|
All programs entered into the Learning Machine Challenge are required to conform to a very simple interface. This interface is console-based, freeing you to use your programming language of choice. In fact, you are free to submit a binary executable if you desire: there is no need for you to reveal your source code to us! This interface facilitates communication with a judge program, which operates according to the rules of the contest games. See 'Official Rules', 'SDK Specs' and 'Protocol Specs' for more details.
When a new game begins, the judge program will inform your player program of the predefined alphabet used by the game. Each move is represented as a string, and the alphabet symbol-set will contain between two and eighty strings. Typically, the alphabet will consist of English lowercase letters. Other games may have more, such as the twenty-six uppercase alphabetic characters, and/or the ten digits.
Once the game is underway, one of three things may happen. Your program may observe a move on its input channel, it may be asked to generate a move on its output channel, or it may be informed of its score. These things may happen in any order, so, for example, your program may observe three moves in a row, be asked to generate two moves, observe one more move, and then be told its score. This sequence will make sense in the context of the game, which, of course, is only known to the judge program.
The LMC2 Protocol allows for the definition of other input channels, such as the Example Channel.
LMC2 consists of two separate phases: Game Submission and Player Submission. While the actual competition is held between players, participants are encouraged to take part in the preceding Game Submission Phase. The tournament will consist of all players in all games, of which there will be between six and twelve. Programs will be ranked by their total score in the tournament, and the program with the highest score will be declared the winner.
Why are we sponsoring it?
Ai Research is a private research project dedicated to achieving the "holy grail" of artificial intelligence: a computer program which has the ability to learn and use a natural language such as English. Our philosophy is that this is achievable using general learning mechanisms, and we embrace the behaviourist black-box approach to learning. We encourage research on state-of-the-art general learning algorithms, and are interested in their strengths and weaknesses with respect to the language learning task.
For more information about Artificial Intelligence Research, please see the Ai home page.
Why should you enter?
The Learning Machine Challenge standardises the interface between a learning algorithm and the outside world. This enables quite different approaches to learning to be compared directly with one another. You may already have a learning system which you have developed. This is a fantastic opportunity for you to prove the abilities of your system to the rest of the world. Can your system beat the best on offer? Can it learn more quickly than a state-of-the-art neural network developed by one of the world's leading universities? These are the questions which the Learning Machine Challenge will answer. We are confident that the results will be surprising and revealing.
Great advances often come from individuals and outsiders who look at the problem from a different angle. You may be a hobbyist programmer with a great idea, and the Learning Machine Challenge gives you the opportunity to try it out. We accept entries from individuals, groups, university departments, commercial research institutes, companies; the list goes on. If you think you have what it takes, there is no