mlthrowaway1953

This is fantastic, thank you!

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

Most employer agreements have fairly draconian claims around IP. And I believe that many universities require their faculty to run consulting, etc. gigs through their legal teams, and the university gets a "right of first refusal" of sorts on the generated IP. Many academics get around this by claiming that they are doing research X at company (more applied, etc.) and research Y at the university (more theory, technically different projects under different grants). But with ML it seems like it would be harder. Regardless, I'd love to figure out if we can get these agreements through freedom of information act requests, to shed light on what state-employees are doing.

How is IP handled in this situation? Especially given the recent discussion [1] on HN about how the recently announced DeepMind Patent Portfolio could be a problem, how is it that state employees (Washington, California) are "co-employed" by an organization to do _exactly_ the research that their home universities work on ? Who owns the IP in this case? Are these sorts of agreements FOIA-able?

[1] https://news.ycombinator.com/item?id=17266951