[mlthrowaway1953]

Is the Fields Medal going to still be relevant in 10 years, when most of the the major mathematical discoveries are made by deep learning and deep reinforcement learning systems? Already systems are learning to reason about concepts [1] and of course there is classical work on proof checkers [2]. It's very likely that the 2028 Fields medal will be awarded to a programmer, not some mathematical super-genius (assuming that the committee is fair, and not biased against machines).

[1] https://arxiv.org/abs/1806.01261 [2] https://en.wikipedia.org/wiki/Four_color_theorem

[cycrutchfield → mlthrowaway1953]

It's shit like this that tells me we deserve another AI Winter. Just at the very least to drive out the dilettantes and charlatans.

[throwaway080383 → mlthrowaway1953]

If you're wondering why you're being down voted, it's because very few people believe mathematical discoveries and proofs are going to be automated any time soon. FCT was exceptional in that the theorists were able to reduce the theoretical proof to a brute force check that no human wanted to do. To be honest, making these claims, especially with such certainty, comes off as rather crank-y.

[pathsjs → mlthrowaway1953]

As a mathematician turned to deep learning, I would say you overstimate the promises of deep learning a bit :-)

[thalesmello → mlthrowaway1953]

Mathematics require the exact type of abstract thinking machines suck at. Machines can execute things fast, learn things fast (provided we have well stablished rules), but than can't (at our current moment in time) come up with useful abstractions to help solve new problems.

So I think it's going to stay relevant.