ICFPC 2012 with crowdsourcing and neural networks
Good day to all.
Everyone is welcome to play a slightly simplified browser version of the famous Supaplex game.
What for? And then, based on your actions, we will then train your bot. But first things first.
')
UPD: Here I will upload the numbers of random cards that you can most likely get:
Simple: 20 , 24 , 32 , 37 , 41 , 43 , 47 , 51 , 56 , 57 , 61 , 62 , 67 , 73 , 77 , 80
Medium: 34 , 36 , 42 , 44 , 59 , 75 , 100
Difficult: 45 , 313
From 12:00 GMT on Friday, the 13th of July, until 12:00 GMT on Monday, the 16th of July, the annual functional programming competition, timed to coincide with International Conference on Functional Programming 2012 (ICFP), takes place.
The contest is notable for many reasons. Personally, I won the 2010 assignment. First of all, the fact that it was necessary not only to program, but also to do a lot of reverse engineering, to break riddles of other teams and create your own, so that they were as difficult as possible to solve.
For those who want to learn more about everything:
For the rest I will tell shortly. It was necessary to invent and upload “machines” to the server and select “fuel” for other people's machines. But not everything is so simple. Initially, even the format in which the answers were received was not known. And the decision of the machines required good knowledge in mathematics.
This year the task is not so funny - you just need to program a robot that extracts lambda from the mine and looks suspiciously similar to the well-known game Supaplex . In order for the participants not to relax, the organizers periodically add new conditions . The competition is divided into two separate offset - blitz (1 day) and the main competition (3 days).
Unfortunately, we found out too late about the contest and could not assemble a full-fledged team. Therefore, our team consists of only two people (up to 20 are allowed, teams less than 4-5 people usually have to be very difficult): me and GreenPeace
First of all, the question was very acute - is it worth it to bother at all? After all, almost nothing will have time. Moreover, no hacking of other people's decisions was envisaged this year, and this was the most fun. Then they decided that it was possible to try to make a robot with random behavior and minimal self-preservation instinct and send it to the blitz test, just for fun. The next thought was to write an emulator of this game and run it by hand to train the neural network. And the logical consequence of this was the transfer of the game to JS and the creation of a site for crowdsourcing . This is even a team of two people. How good this will bring results - we'll see later.
Time is short, there are no servers at hand, so the static lies on the people (I hope it doesn’t have any habraeffect), and the minimum required dynamics is on the first free hosting.
The task of the robot is to collect all the lambda and go to the elevator. For each collected lambda robot gets 25 points. If he brings them to the elevator, another 50 each. If you stop mining (abort - button "A") - 25 each. Well, if you drown or get a stone on the head - then the robot will not receive any additional points. For each step, the robot loses 1 point.
In addition, the stones fall, if there is where, roll with each other (right or left) and with lambda (only right).
A few hours after the start, the organizers added a flood. Once a few moves the water rises, the robot can dive into it, but not longer than several moves. How much is this very “several” depends on the card.
Detailed rules in Russian can be found in the next post .
Currently 999 maps are available for the game (15 cards from the organizers and 984 randomly generated). Cards from the organizers are made by hand, interesting and obviously passable. For random cards, we do not guarantee anything. On some, nothing at all can be typed. On others, the best strategy is to collect available lambdas and stop the game. In addition, there are no floods on random maps. But this will be corrected immediately after our team will sleep a little. Since the beginning of the contest, 16 hours have passed and now is the time to take a break for sleep.
Planned updates for the game:
Questions, requests, bug reports write in comments or in private messages. All playing - many thanks.
Everyone is welcome to play a slightly simplified browser version of the famous Supaplex game.
What for? And then, based on your actions, we will then train your bot. But first things first.
')
UPD: Here I will upload the numbers of random cards that you can most likely get:
Simple: 20 , 24 , 32 , 37 , 41 , 43 , 47 , 51 , 56 , 57 , 61 , 62 , 67 , 73 , 77 , 80
Medium: 34 , 36 , 42 , 44 , 59 , 75 , 100
Difficult: 45 , 313
A bit of history
From 12:00 GMT on Friday, the 13th of July, until 12:00 GMT on Monday, the 16th of July, the annual functional programming competition, timed to coincide with International Conference on Functional Programming 2012 (ICFP), takes place.
The contest is notable for many reasons. Personally, I won the 2010 assignment. First of all, the fact that it was necessary not only to program, but also to do a lot of reverse engineering, to break riddles of other teams and create your own, so that they were as difficult as possible to solve.
For those who want to learn more about everything:
For the rest I will tell shortly. It was necessary to invent and upload “machines” to the server and select “fuel” for other people's machines. But not everything is so simple. Initially, even the format in which the answers were received was not known. And the decision of the machines required good knowledge in mathematics.
This year
This year the task is not so funny - you just need to program a robot that extracts lambda from the mine and looks suspiciously similar to the well-known game Supaplex . In order for the participants not to relax, the organizers periodically add new conditions . The competition is divided into two separate offset - blitz (1 day) and the main competition (3 days).
Unfortunately, we found out too late about the contest and could not assemble a full-fledged team. Therefore, our team consists of only two people (up to 20 are allowed, teams less than 4-5 people usually have to be very difficult): me and GreenPeace
First of all, the question was very acute - is it worth it to bother at all? After all, almost nothing will have time. Moreover, no hacking of other people's decisions was envisaged this year, and this was the most fun. Then they decided that it was possible to try to make a robot with random behavior and minimal self-preservation instinct and send it to the blitz test, just for fun. The next thought was to write an emulator of this game and run it by hand to train the neural network. And the logical consequence of this was the transfer of the game to JS and the creation of a site for crowdsourcing . This is even a team of two people. How good this will bring results - we'll see later.
Time is short, there are no servers at hand, so the static lies on the people (I hope it doesn’t have any habraeffect), and the minimum required dynamics is on the first free hosting.
About the game
The task of the robot is to collect all the lambda and go to the elevator. For each collected lambda robot gets 25 points. If he brings them to the elevator, another 50 each. If you stop mining (abort - button "A") - 25 each. Well, if you drown or get a stone on the head - then the robot will not receive any additional points. For each step, the robot loses 1 point.
In addition, the stones fall, if there is where, roll with each other (right or left) and with lambda (only right).
A few hours after the start, the organizers added a flood. Once a few moves the water rises, the robot can dive into it, but not longer than several moves. How much is this very “several” depends on the card.
Detailed rules in Russian can be found in the next post .
Currently 999 maps are available for the game (15 cards from the organizers and 984 randomly generated). Cards from the organizers are made by hand, interesting and obviously passable. For random cards, we do not guarantee anything. On some, nothing at all can be typed. On others, the best strategy is to collect available lambdas and stop the game. In addition, there are no floods on random maps. But this will be corrected immediately after our team will sleep a little. Since the beginning of the contest, 16 hours have passed and now is the time to take a break for sleep.
Planned updates for the game:
opportunity undodone- cards with flooding
- improved card generation algorithm
direct links to mapsdone
Questions, requests, bug reports write in comments or in private messages. All playing - many thanks.
Source: https://habr.com/ru/post/147798/
All Articles