https://www.euro-online.org/enog/inoc2007/Papers/mac-slots.html https://www.euro-online.org/enog/inoc2007/Papers/m https://www.euro-online.org/enog/inoc2007/Papers/mac-slots.html

New professorship combining Chemometrics and Cybernetics: Professor/Associate Professor in Big Data Cybernetics

ianm's picture
Forums: 

https://www.jobbnorge.no/en/available-jobs/job/127389/professor-associate-professor-in-big-data-cybernetics

A new professorship/associate-professorship is now open  in the Department of Engineering Cybernetics at Norway’s biggest university, the Norwegian University of Science and Technology (NTNU), Trondheim, on the new combination of very high-dimensional chemometrics (megavariate soft data modelling) and control theory/nonlinear dynamic modelling.

This is an opportunity for chemometricians to contribute to science and society in new ways, and to learn a lot from other sciences.

Younger chemometricians, not yet qualified for a full professorship, may also apply!

Trondheim is a great city to live in – with lots of old history, a beautiful fjord,  nice hills and mountains nearby, good international communications, with a high academic profile  and an intense, entrepreneurial high-tech culture.

Application deadline: 2016-09-30.

Chemometrics in  Norway is taking on a new—and for me surprising—flavor these days: in electronics & control theory, of all things! I think it is a lot of fun, even though it is quite a mouthful for me, being trained as a biochemist and sausage scientist.

What I and my cybernetics colleagues want to develop—in cooperation with colleagues internationally—is an academic discipline which prepares science, technology and society at large for better use of the deluge of quantitative, but nonselective measurement data that increasingly becomes available. The pragmatic Chemometrics and NIR spectroscopy cultures, with their powerful tools to discover and describe both known and unknown phenomena, have a lot to offer in this context! We need to combine data-modelling and mechanistic-modelling methods from different disciplines.

Any multivariate method may be useful, as long as it catches relevant and reliable patterns in data, is easy to validate statistically and easy to interpret graphically. That includes chemometric methods as well as methods from control theory. In addition, we want to take inspiration from—and work with—the statistics & machine learning communities.

But we do want to avoid

1) “Macho mathematics”:  overly theoretical approaches to mechanistic modelling and statistical distribution theory

2) “Gucci statistics”: superficial significance testing with high risk of false discovery and

3) “Blind machine learning”: alienating “black box” ANN solutions that are difficult to interpret in light of prior knowledge.

We want to develop a science culture where Big Data makes people smarter, not dumber. We want  to develop a technology culture that solves theoretical and practical problems by combining mathematically compact modelling, statistically valid assessment, cognitively adapted graphics—but also a radically better ability to let the real world—via Big Data—inform us about all the phenomena out there that we still do not understand.

If you know anyone who could be interested in working with my colleagues and me at Norway’s biggest university,  in shaping this cross-interdisciplinary field—please pass the information on to them!

Harald Martens

Prof. II, Dept. Engineering Cybernetics, NTNU, Trondheim [email protected] Mobile +47 95075025

(posted on behalf of Harald Martens by Ian Michael)