3 ms·
Are we witnessing the dawn of post-theory science?
- jleyank 5y agoCatchy title. Empirical models, while useful, can't explain the how or why of a system. Having machine-learning (err, AI) models is great, but there's no idea of what will happen with edge cases or novel situations. Such models can provide data which augments experimental data. Theories, ideally, can be used to propose new experiments. Data-only models can only interpolate.
- pella 5y agodups: https://news.ycombinator.com/item?id=29862382 https://news.ycombinator.com/item?id=29862382