For a research to have a substantial influence, it must either be experimentally or mathematically proven. For example, you have to either fabricate a chip and board for experimental verification, or come up with a mathematical model to prove that your method is second to none. However, most of the recent works by the students only feature imitating experiment and simulation. While you can easily obtain results with these, it does not have remarkable influence on the academic world nor the industry. Therefore, I suggest that the KAIST students take a lot of courses of mathematical science department. The more math you know, the larger part of the world you can model. And the more you can model, the better you can engineer.