How to Do Core Work - Minimizing Bureaucracy

Aims: Promoting doing good research - not politically motivated research for tenure and promotion or accolades


Core work is defined to be foundational research, which enables the development of superior products and/or services. This definition applies to teaching as well. This conclusion is an outgrowth of a two-day sidebar held with researches at CMU at an applied AI conference. When you stop and think about it, why do we do what we do? Is it to get paid, because we are good at it, because we serve progress, because it motivates us, because it serves to increase the gross GDP, because it pays for our health care, etc.? We can all agree that it is for all of the above reasons. But, it is equally clear that we do not do it for lack of anything better to do. This moves the question from the realm of the "why" to the realm of the "how". How do we accentuate the positive and eliminate the negative from our work? It is said that the first step in solving any problem comes from a definition of that problem. Most of us, like the frog in a pot of water being slowly brought to a boil are simply unaware of our developing situation. We need to ask ourselves the "what if" questions. We need to help minimize bureaucratic waste of time and promote original R&D - i.e., not publishing for tenure or promotion - publishing because you have something important to say., which is defined to be worth other researchers time to read. Otherwise, we waste scarce financial resources, which serves no one. More and more research these days wastes authors and readers time. The application changes, but the science and engineering have little to nothing to offer to make the work worth reading. In KAIS, we foster an environment to promote true innovation - from new neural architectures to engineering more energy efficient photonic computers and the acquisition of heuristics as competition for quantum computation and the like. We strive for novelty in all that we do because that defines progress and all TCs need serve the same goal to be true to their charters.

For example, if the calculus were to be invented today, how many innovative theorems would be defined in comparison to analytical studies? I can assure you that good work takes time and is not served by the pressure to publish. It took 300 years for calculus to evolve to where we find it today.


Drop impact factor journal ratings and counting professor publications. Instead define professor's contributions by the qualitative meaning of their top 5-10 works, what groundwork they lay, and what future they enable. Otherwise, we stymie innovation, waste taxpayer dollars, and in the end unnecessarily limit the growth and dissemination of good science and engineering. We may be responsible for a large portion of funding cuts. You cannot have technical progress in the absence of a foundation of mutual trust and respect - between professors, their students, their administrators, and their sponsors. That is not the way it is now; but, that is the way it needs to become if progress is to be best served.

About the Authors:

Chairs of the TC Knowledge Acquisition in Intelligent Systems (KAIS)

"Dr. Dr. Shu-Ching Chen

Dr. Stuart Rubin

Dr. Shu-Ching Chen