28 Apr
2020
28 Apr
'20
4:51 p.m.
Don't recall this getting discussed here: https://arxiv.org/pdf/1912.01412.pdf Interestingly treats symbolic integration as a kind of linguistic translation. Leverages the cost asymmetry "trap door" versus differentiation when training. Thoughts? (Might something similar deconvolve a long digit string into two prime factors? Multiplying up a training set of a couple of hundred million products is easy!)