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 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.
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.
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
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!
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
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 😁.
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...
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.
