A chat with a certain Witek Nazarewicz on uncertainty quantification

By Colin Adcock on November 10, 2015 • ( 1 )

Witek Nazarewicz (Michigan State University) is one of the most highly recognised scientists in nuclear physics today, with over 380 scientific works to his name and a h-index of over 70. Recently he’s been encouraging experimentalists and theorists to work more closely together, and is a co-organizer (together with David Ireland and Bartek Szpak) an upcoming conference at ECT* in Trento, Italy.

He’s also a Guest Editor for the JPhysG focus issue ‘Enhancing the interaction between nuclear experiment and theory through information and statistics‘, which we wrote about a short while ago. We asked him a few questions.

What are you and your team currently working on?

The main area of my professional activity is the theoretical description of atomic nuclei, including exotic, short-lived nuclei that inhabit remote regions of nuclear landscape. This research invites a strong interaction between nuclear physics, many-body science, and high-performance computing. Our goals are manyfold. We want to develop next-generation nuclear interactions that would provide quantitative description of nuclei, their decays, and reactions, and also neutron star structure. We are working on microscopic frameworks to describe various nuclear collective motions, including fission. We also struggle to unify structure and reaction aspects of nuclei.

You are a leader of the ISNET initiative (Information and Statistics in Nuclear Experiment and Theory). What is it, and why is it important?

Exact models are seldom available in realistic nuclear modeling. At the heart of ISNET is the uncertainty quantification of theoretical predictions. The basic question is: what is the best way to use experimental data in the formulation of theoretical models that attempt to explain the results of experiments and make predictions for new observables, often involving huge extrapolations? To this end, modern nuclear theory is developing tools to deliver uncertainty quantification and error analysis for theoretical studies as well as for the assessment of new experimental data.

How can we improve the way our communities interact?

The importance of interdisciplinary meetings, involving a broad audience (theory and experiment, various areas of science, …) cannot be overstated. The major theoretical nuclear theory centres, such as ECT* in Trento or the Institute for Nuclear Theory in Seattle, are great facilitators of inter-community interactions. Easily accessible websites, such as the ISNET portal at JPhysG, are great communication vehicles for spreading the news and excitement.

A schematic illustration of systematic (accuracy) and statistical (precision) uncertainties of nuclear models.

Tell us about the upcoming ISNET workshop. What do you hope to achieve?

The unifying theme of the upcoming ISNET conference is to illustrate the extent to which uncertainty is a key quantity in nuclear modeling, and to showcase applications of the latest methodologies. Through presentations and discussions, we want to increase the awareness of the related issues in the nuclear community.

The ISNET workshop is being held from 16-20 November 2015 at ECT*, Trento, Italy. Web: gamma.ifj.edu.pl/isnet

Do you think other scientific communities could benefit from this type of effort?

While in some areas (such as atomic physics), theoretical results are routinely presented with uncertainty, this is not a generally accepted principle. The tools of uncertainty quantifications are interdisciplinary, and different communities can benefit by joining ISNET-type discussions.

Anything else you would like to add?

The train has left the station. The need for uncertainty estimates of theoretical models has been recognized in the nuclear physics community. So the question is not whether to do it or not, but how to do it best.

This work is licensed under a Creative Commons Attribution 3.0 Unported License