4 ms·
I'm happy to notice that mathematics is no longer confined to academic journals. Anyone can contribute at any level. I wish a similar destruction of academic wa
by nyc111 8y ago
I'm happy to notice that mathematics is no longer confined to academic journals. Anyone can contribute at any level. I wish a similar destruction of academic walls to other fields, especially physics.
- anonytrary 8y agoI'd imagine it being the possible for theoretical physics, as it is in math. Otherwise, one problem is when you start needing advanced equipment and laboratories. The barrier to contributing is higher.
- n4r9 8y agoThis case is exceptional in that the problem and solution can both be explained in fairly elementary terms. Combinatorics is really the only field of mathematics with many such opportunities available.
- lasagnaphil 8y agoI think computer science is another field that individuals can potentially break through the academic wall, because you would only need a decent desktop to do your research... (I think quite a lot of resarchers do stuff in AI, graphics, etc using gaming-level PCs instead of huge workstations/servers)
- blt 8y agoThis is true in general, but deep learning with lots of hyperparameters gives an advantage to organizations with clusters that can run a huge hyperparameter search quickly.
- Zephyr314 8y agoSigOpt (YC W15) has an academic program that allows full access to their optimization platform for academic use [1]. Hundreds of academics around the world have used it and there are already dozens of papers that have been published that have benefited from it [2]. These techniques can be orders of magnitude more efficient than a standard hyperparameter search and really cut down the barrier to entry for these types of results. [1]: https://sigopt.com/solution/for-academia/ https://sigopt.com/solution/for-academia/ [2]: https://sigopt.com/research/publications/ https://sigopt.com/research/publications/