New working notes: "Reduced MNIST: how well can machines learn from small data?" http://cognitivemedium.com/rmnist
Show this thread
-
If I've understood q correctly, Brenden Lake, Josh Tennebaum et al have been working on this for a while, they proposed MNIST Transpose (few examples for 1000s of categories :) back in 2011 (if not earlier) http://web.mit.edu/jgross/Public/lake_etal_cogsci2011.pdf … https://cims.nyu.edu/~brenden/LakeEtAlNips2013.pdf … https://arxiv.org/abs/1604.00289
Oh, there's no claim at all the problem is original, it's meant as a good playground. Pretty sure people were working on this (well, not this specific variant) back into the 70s, if not earlier.
-
-
That's a nice paper by Tenenbaum et al, thanks!
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
oh i wasn't trying to suggest that you were suggesting that it was original :) I thought maybe the 'alien alphabet' one shot learning you were looking for might be the tenenbaum one
-
Ah, I misunderstood! Thanks for clarifying. And thanks again for pointing out the paper, it looks really nice.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.