(Back to the menu - click here.)

“Machine learning and AI for the sciences -- towards understanding”

Date:

Download-files:

Time:

Thursday, 08. June 2017

Video-Recording for any system with MP4-support

- Video.mp4 (ca.330 Mb)

15:15 – 16:15

Abstract :

The seemingly chaotic world around us often exhibits remarkable regularities and

repetitive patterns, suggesting a simple origin. That in turn can point to the relevance

of symmetries in nature, often reflecting a geometric underpinning.

Mathematically, the exploitation of symmetries frequently involves algebraic techniques.

In atomic nuclei, the use of structural symmetries or algebraic approaches is one of the

most important tools for the study and understanding of nuclear structure. An example

is the Wigner supermultiplet symmetry SU(4) or the symmetry SU(3), pioneered by Elliott

in the 1950’s to study deformed light mass nuclei in the s-d shell. In heavy nuclei, this

symmetry breaks but a well-known approximate symmetry, pseudo- SU(3), has been

actively studied for many years. Since the 1970’s, the interacting boson model – based on

an algebraic group theoretical foundation – and its offshoots and extensions, has provided

a very successful approach to understanding the structure of myriads of medium mass and

heavy nuclei. Algebraic methods often provide simple, analytic, and parameter-efficient

approaches to the structure of many-body systems.

This talk will provide a selected perspective on some of these and then turn to a new entry

in this field, an approximate symmetry called proxy-SU(3).

Proxy-SU(3) is motivated by a consideration of the spatial overlaps of certain nucleon

wave functions, in particular for deformed nuclei. Like pseudo-SU(3), it exploits a specific

orbit substitution to attain a space that can be described by an SU(3) symmetry.

This in turn allows a number of parameter-free predictions of the structure of heavy

deformed nuclei using the simplest analytic expressions. We will introduce this symmetry,

show the nature and impact of the approximations involved, and outline how its irreps

can be used to make specific predictions of the deformation variables  and , of prolate

dominance in deformed nuclei, and of the locus of the prolate-oblate transition region.

Overall, the agreement is good, but, at the same time, specific classes of discrepancies point

to the need for improvements to the model that take into account missing ingredients.

The future colloquium program can be found at: