Thursday 17 Aug 1pm AEST 

Click here  to watch the recording on the AIP YouTube channel.

Abstract: Scattering is described in physics by the relations between asymptotically ingoing and outgoing states. The corresponding notions in Einstein’s theory of gravity are the past and future light-like infinities as introduced by Roger Penrose. They rely on the conformal structure of the Lorentz manifold describing the system. In this talk, we will discuss how conformal methods can be used to describe the scattering of gravitational waves geometrically and numerically.

Thursday 13 July 1pm AEST 

Click here  to watch the recording on the AIP YouTube channel.

Abstract: Dark matter is an elusive substance which, despite considerable effort, continues to evade detection. In this talk we will tour through the realms of particle, nuclear, atomic, and astro-physics, surveying the rich interdisciplinary research that propels our search for dark matter. The efforts of astrophysics and cosmology provide us with the necessary cosmic context, while the fields of nuclear and atomic physics ground us, informing and guiding the search for dark matter in the laboratory. We will explore the compelling evidence for the existence of dark matter and how we can best determine its true nature. 

Peter Drummond, Margaret Reid, Alex Dellios, Bogdan Opanchuk, Simon Kiesewetter, and Run-Yan Teh

Thursday 27 April 1pm AEST

Click here  to watch the recording on the AIP YouTube channel.

Abstract:   There are experimental claims of computational advantage on quantum computers. This raises  theoretical questions of validation for the random-number generation tasks that are solved. How does one verify the output? Are the answers obtained even correct, and how can one test this in practise?

Brute-force computational verification is not possible. No classical computer is large or fast enough for this, without taking billions of years. Even computing the distributions is exponentially hard, not just from time and memory, but also due to precision constraints, as there is insufficient precision.

For Gaussian boson sampling tasks, we show that simulations in quantum phase-space can solve this, by generating any diagnostic that is measurable. This uses an FFT binning algorithm to obtain computable statistics, with up to 16,000 qubits in large test cases, far larger than in any experiment.

The result is that recent experimental data from China and USA is significantly different from theory, with over 100 standard deviations of discrepancy for some measured output statistics. Possible explanations are explored, but this is a nontrivial physics problem, and we do not have a complete explanation.

This does not disprove the computational advantage claims. These are very hard tasks, and we do not directly generate the required numbers. However, the outputs do not survive the chi-squared tests one would normally use to test validity of random numbers, as used in numerous cryptography applications.

Finally, we point out how similar techniques may in future be useful in testing other quantum network and computer designs. The principle is to use scalable methods that generate probabilities, rather than trying to use naive algorithms on classical machines, which are now totally impractical.

Thursday 6 April 11am AEST

Abstract:   Our understanding of the structure of matter, encapsulated in the Standard Model of particle physics, is that protons, neutrons, and nuclei emerge dynamically from the interactions of underlying quark and gluon degrees of freedom. I will describe how first-principles theory calculations have given us new insights into this structure, including recent predictions of the contributions of gluons to the pressure and shear distributions in the proton, which will be measurable at the planned Electron-Ion Collider. I will also discuss studies of light nuclei which provide insights relevant to dark matter direct detection experiments and to searches for evidence of the Majorana nature of neutrinos through neutrinoless double beta decay. Finally, I will explain how provably exact machine learning algorithms are providing new possibilities in this field.

Thursday 6 Oct 1pm AEST

Click here  to watch the recording on the AIP YouTube channel.

Abstract:  A strong coupling between light and matter can be achieved by embedding two-dimensional layers of semiconductor in an optical microcavity. This results in the formation of exciton-polaritons, which are hybrid part-light, part-matter particles that are capable of realising a range of quantum many-body phenomena. However, the interactions between such polaritons are still not well understood despite their fundamental role. In this talk, I will discuss recent theoretical progress in understanding the microscopic properties of polaritons. In particular, I will show how the two-dimensional geometry plays an important role and leads to highly counterintuitive results, such as light-enhanced polariton-polariton interactions.

Thursday 15 Sept 1pm AEST

Abstract:  Einstein, Podolsky and Rosen (EPR) presented an argument that quantum mechanics is an incomplete theory. However, the argument assumes local realism which is falsifiable by Bell’s theorem. Here, we re-examine the argument, by presenting a mapping between microscopic and macroscopic Bell tests.  The macroscopic tests involve qubits based on two macroscopically-distinct coherent states and suitable unitary interactions. This compels us to address how the macro-Bell tests can be compatible with the important concept of macroscopic realism. We show that deterministic macroscopic realism is falsified by the macro-Bell tests, and therefore define the weaker assumption that we call “weak macroscopic realism”, which takes into account the dynamics associated with the choice of measurement setting. We show consistency of weak macroscopic realism with the Bell violations, as well as macroscopic versions of Greenberger-Horne-Zeilinger, Wigner’s friend and delayed-choice experiments. This brings us to deduce a macroscopic version of the EPR paradox based on weak macroscopic realism, thereby re-opening the question of the incompleteness of quantum mechanics. We then examine the measurement problem by proposing a model for measurement using simultaneous forward- and backward-propagating equations in time, derived from Q function dynamics. We demonstrate a causal consistency, and distinguish measurable from unobservable variables, which leads to models of realism and causal relations involving loops. We show that the new model supports weak macroscopic realism and explain how consistency with EPR-Bell correlations can be achieved.