ฝัง
- เผยแพร่เมื่อ 26 ก.ย. 2024
- Top 3 games of new GeoGuessr AI so far (new high score)!
More on the development of the AI (how the AI learns to play):
• AI learns GeoGuessr an...
Subscribe for more AI adventures:
/ @traversedtv
Join the discord:
/ discord
i need a rematch
When are you free? 😎🤖
this would be a great video
Set it uppppp
I could have expected Rainbolt commenting, but Rainbolt Two!! Yooooo
New vid?
geoguesser ai is looking much more refined! i love how you show the heatmap for what the ai is thinking. that is very interesting.
I totally agree! So much more interesting to watch the guesses with the heat map! I'd also love to zoom into the relevant area for the predictions but didn't have time for that yet 😁. That would make it more similar to how geoguessr shows results too.
I'd be very interested to see a heat map of what parts of the image the AI extracts important features from. Maybe you could build it by blocking out areas of the image and see how it affects the prediction. (There's probably a better way to do that I haven't studied this since college) it would be dope if the AI discovered new meta that humans never found
Amazing, i love it, i have one thing though, maybe you should switch training map to "A diverse world"? "World" has much higher chances of getting countries like USA, Japan, Russia, UK, etc. "A diverse world" map is more balanced and gives much more equal chances of getting the rest of the countries.
it could just learn literally every, location on adw and 5k all the time because there are so few
Very cool! What would be super interesting to see is where the AI looks in the images to make the guesses. Then we could maybe learn from the AI
It would be really interesting to do some interpretability analysis of the network to see which features it is focusing on to make its guess. Could reveal tricks for human players to use!
Totally agree! I tried that but the results where very inconclusive... It's a bit tricky due to the relatively high image resolution. Will try again when I got time!
Is the current network only based on the CNN? I'm always a big fan of combining the strengths of AI and classic methods. Can you also make a feature extraction and use that as input? This would be something like: we're driving on the left, so it can only be one of those countries.
Sure, with enough data all of that can be learned, but maybe a few well selected features could cut down the amount of needed training data significantly.
Nice job btw. :). This is my favorite GeoGuesser AI so far!
You should probably have the AI play on a diverse world, as the default world map is highly skewed towards capital cities.
but there aren’t enough locations on adw so the ai would just know all the locations. i guess you could train it on 1 map and then play it on another map though, to see how it translates over
Rainbolt getting a run for his money soon I can feel :)
This score with 3 America and 1 Japan seeds is scary
Very scary. A government could take a photo from your photo and instantly know your location within cell phone tower range...
Thus even with your location service and data turned off, you could be found by police or military...
I know nothing about this, but I have a question. How can you use pretrained network and make it work? The network would have to be trained for the exact same purpose, wouldn't it?
Also about country codes: you meant to say output, not input, right?
You can also use models that are trained for a different purpose. It requires some effort though: in simple words you remove the last step the AI would usually do / the last layer of the network. Then you add a new layer and only train tust new layer. This allows the network to use rather general knowledge on images from the old layers while learning how to solve the new problem (geo-location) with the new layer.
And yes: output is what I meant 😁.
Would there be any way for us to help donate some computing power into training the AI?
Also, your discord link is invalid btw, you should update it
Hi Zabrid,
thank you for the offer! Currently it's fine, but I'll see how that develops.
I updated the discord link 👍
This is so cool!
Finde ich persönlich super interessant ;)
Danke dir ;)
Maybe there's some complication I'm not seeing, but would it not be easier to make it a regression problem in the coordinates rather than dividing the map into sections?
Being a big fan of regression, I'm thinking about that too, but I would assume that unless the model is really well trainined and already nails the region, the regression would be much less accurate. Take an example in which the model thinks of a place in the US and one in Europe, the regression model would probably end up somewhere in Iceland. Maybe a regression on the cluster of most certain guesses I'm the heat map might make sense, but it could already be doing that.
very nice progress! the discord link seems to be invalid or expired though
Thank you! The link was inspired indeed!😅
Wäre diese AI nicht perfekt, um Bilder ohne Metadaten zu tracken?
Dumb question if I’m wrong, but are you by chance my marketing Professor?
Supercool project - but can I ask, what’s its average score at the Moment?
Nope, I'm not a professor 😁! The average score is around 16.600 atm. But I didn't work on it for a while.
can you also make an AI trained on ADW and ACW?
Gibt es Pläne auch andere Maps zum Training zu verwenden?
Ich hatte genau das gleich vor 😂😂😂
Damn it's getting goated.
Gute Arbeit!
Danke!
Bist du deutsch?
Ja ;)
Haha nice
is there a download?
Unfortunately not. I might do a few AI vs viewer duels on discord soon.
@@TraversedTV what program do you use to code the ai?
@@zerxioo A bit of Javascript for the brower plugin and python for the AI itself.
Du klingst deutsch
Das stimmt 😄
Personally, I would find it cool to see how the image recognition works and some of the features/patterns it uses to find the location. Nice work!