Hi all - looking for some feedback. A friend of mine runs Arkadium, the largest online gaming platform in the world. He has 20M users / month across 100 websites (all major newspapers use his platform). Recently he was contacted by one of the large AI companies who wanted to purchase game data from him. He has both JSON and images. Games are bridge, backgammon, mahjong, sudoku, poker etc. He asked me if I could if I could investigate as he’s in New York and I live in San Francisco. Any thoughts? Data is 100% human created and never been used for training (except by the company that contacted him out of nowhere - they ust started using it).
1 Like
Hi!
Maybe this is of interest in this context.
There is a submit interest button at the bottom of the page.
1 Like
I’d imagine so. I’d say it depends on what the data is and what is it used for, but for example if the data has all the different moves in a Backgammon game and you have lots of data then that can be used to start training an AI to play Backgammon. Then extend that to each of the different games. As I say it would of course depend on what is in the data, so for example in the Poker data it would need to know what cards were public and what cards were known only to each player so that the AI could take that into consideration.
Data is JSON text an images of the games. About 80M games / month from 20M users - Crossword, poker, hex, hearts, Euchre, Gin Rummy, Bridge, Mahjong. I’m wondering if this has applications in AGI? Limited information games can be extended to other real world issues such as stock trading. Thanks for the answer.