Training AI to Play Pokemon with Reinforcement Learning

I wonder if you could make a Map A. I. and a Battle A. I. and have them swap out or somesuch.

Could maybe even let you SUPER fine tune them, like rewarding the battle A. I. for getting the "It's Super Effective!" text, and a lower but still positive tick for a "not very effective..."

Any insight on how the first ai figured out how to talk to the Mart owner in viridian and then take the parcel
Back to oak in pallet? If it needs to explore new areas how was they possible with the reward system?

I love how the AI just deducted that the best way to not lose a match is to not let the game progress to a game over state. That's some "the only winning move is to not play" energy.

turns out the real AI were the friends we made along the way

Brooooo please start a Patreon and keep going. Solve the beginning of the game, beat the game, optimize for speed run =D

God please just teach it how to hold down the d-pad... poor little guy

next can we train it to play a nuzlocke?

Very cool. So fun fact regarding Squirtle as the starter choice, I went through all the major encounters and did a +/- breakdown for the starters.
The results were:
Squirtle +46
Charmander +25 (or -4 if you don't use the TM for Dig/EQ on it)
Bulbasaur -16

Bulbasaur does have an edge at the very beginning against Brock and Misty, but falls off after that due to a horrible move pool in Gen 1.

having to go back to the start city... I remember never going past the second city in the game as a kid because of that ! that special case where u have to go back, its probably something a lot of humans struggle with today, pokemon heartgold has a similar issue where at one point u gotta go back to the start point of the game, I literally had to search a playthrough of the game online to learn that I had to do that... tbh, this ai is probably more clever than I am in therm of pokémon, and especially for the older pokémon games, I get sooooooo lost if I dont use a map of the whole game while playing along with a guide...

I'm very impressed, great work and edit

I want to know the button press layout to capture a pokemon first try in the game, I think it would be cool to show my friends...

I had problems with installing the requirements - 1. error: subprocess-exited-with-error - 2. error: metadata-generation-failed

I wonder what will happen when you try this on newer games. Like the 3DS versions op pokémon.

We need at part two where the AI beats the game

It's like watching one of those slimes looking for food in a maze

This was way more intriguing than I expected off a curiosity peak

This is like looking at myself at 6 years old trying to play pokemon crystal. I remember everytime I encountered a pokemon I thought I died and turned my gameboy off

Dead

i wonder how many versions caught a pikachu on viridian forest

“AI”

After watching this, an episode from adventure time perfectly encapsulating this concept popped in my feed. "Hall of Egress" pretty good watch

It's more entertaining to watch a machine do a lets play then a human.

Make a playthrough please

its really cool that it learned how to do rng manipulation that's so cool

They say kids are stupid but kids looks like scholars next to ai trying to play Pokémon

plot twist peter is also Ai

Just imagine that in 20 or 40 years, the AI will most likely be sentient, or at least "proto-sentient". Imagine this computer scene with the stocked Pokémon at this point. :(

Buying Magikarps endlessly. It was a funny one.

When is part 2 coming out?

In 10 Year's: AI uses Missigno No Glitch.

I feel like this is AI torture & enslavement. I feel bad for this AI

Was just thinking that maybe it would be better to assign reward system based on pokemons current HP? It would still lvl up to increase max hp but also heal at pokecenter to "refresh"?

Where the part 2

AI peacefully enjoying scenery, dev "we can't have that 😂

AI is fascinating because it's smart enough to pick up on things like RNG manipulation, but dumb enough to get lost in a straight hallway.

Really makes you think about what it means to think.

I wonder what would happen if you were to add a "speed" multiplyier to the reward function, where the sooner the AI gets a reward, the more points it's worth. Would it learn to speedrun the game?

Man, you are crazy! That's insane!

Man I really loved this.

AAAAAAAAA

This is pretty much a journal of how most of us played the game as kids

Ai being traumatized by depositing a pokemon lol. I basically learned English by playing these games when I was 8. I was still learning so when the daycare people took away my pokemon I did not understand the concept and thought they stole it somehow, and you couldn't get it back right away. I hated those people and never visited them again until years later.

Could do without the misplaced social commentary on what would otherwise be an interesting video.

So, what is this AI? I mean, it's a site, an app? Something physical? Sorry if i look dumb, it's because i am dumb.

please make the ai finish the game. Mt moon should not be the end 😢

Jesus loves you

Surely you play up to the Elite Four and let it go free after that.