What’s the hardest video game you’ve ever played? If it weren’t for it QWOP then let me tell you right you know that you don’t know how difficult a game can be. The simple deceptive running game is so difficult to master that even one You trained using machine learning still only collected a top 10 score instead of breaking the record.
If you’ve never done it played QWOP before, you owe it to yourself try and see if you can even get your sprinter off the starting line. Developed by Bennett Foddy in 2008, QWOP was inspired by an ’80s arcade game called Track field that requires players to mindlessly crushing the buttons to win a race. QWOP takes a different approach and instead uses the players four keys to control the individual movements of a runner’s thighs and calves-A runner who behaves like a floppy cloth doll and is submissiveworld physics, including the effects of gravity. It may seem simple, but mastering the timing and cadence of the keystrokes needed for the sprinter to move forward strangely can be incredibly frustrating.
Wesley Liao was curious how good an instrument like AI is he was trained to do things like that realistic animation old photos of the deceased loved ones, would play QWOP. After first creating a Javascript adapter that would allow an AI tool to actually play and interact with the game, Liao’s first machine learning attempt made AI play the game on its own and learn which actions led to positive results. (sprinter going forward and increase its speed) and which of these led to negative results (bending of the sprinter’s trunk too close to the ground.) With this approach, AI learned a “knee scraping” technique that would have managed to make it go over 100-the finish line of the meter, but not at recording speeds.
Liao’s next attempt to train an AI model recorded videos of their game, trying to succeed in the game, including the use of longer leg steps, which are crucial for speeding up and crossing the finish line with a decent time. The approach was a bit more successful, but the AI failed to master a special technique used by Advanced QWOP players involving an upward swing of the legs to generate additional momentum.
G / O Media may receive a commission
Finally, Liao contacted a veteran player known as Kurodo (@cld_el on Twitter), one of the top QWOP speed runners from around the world, who recorded 50 videos with themselves playing the game at the expert level. But even with access to the best possible game techniques, Liao found that the best results came from a machine learning training regime that involved 25 hours of AI play on its own, 15 hours of learning. from data collected from expert Kurodo races and another 25 hours of self-play.
But even with all those efforts, QWOP-Playing the best 100 AIthe result of the subway line made it to the finish line in 1 minute and 8 seconds –a top 10 completion. Conformable Speedrun.com, the current world record of 100 meters per line is only 48 seconds, set just a month ago. Liao is confident with more training and another reward system (the way AI finds out he did something right), setting a QWOP the world record could eventually happen, though, because it’s a computer that plays the game, the record may never be officially recognized.