Quantum Foundations

What can machine learning teach us about quantum mechanics? I will begin with a brief overview of attempts to bring together the two fields, and the insights this may yield. I will then focus on one particular framework, Projective Simulation, which describes physics-based agents that are capable of learning by themselves. These agents can serve as toy models for studying a wide variety of phenomena, as I will illustrate with examples from quantum experiments and biology.

We construct a hidden variable model for the EPR correlations using a Restricted Boltzmann Machine. The model reproduces the expected correlations and thus violates the Bell inequality, as required by Bell's theorem. Unlike most hidden-variable models, this model does not violate the locality assumption in Bell's argument. Rather, it violates measurement independence, albeit in a decidedly non-conspiratorial way.

In this talk I am going to describe Spekkens’ toy model, a non-contextual hidden variable model with an epistemic restriction, a constraint on what an observer can know about reality. The aim of the model, developed for continuous and discrete prime degrees of freedom, is to advocate the epistemic view of quantum theory, where quantum states are states of incomplete knowledge about a deeper underlying reality. In spite of its classical flavour, many aspects that were thought to belong only to quantum mechanics can be reproduced in the model. 

I am going to describe our results regarding the formulation of rules for the update of states after measurement. I will do it for systems of discrete prime dimensions and I will then give the idea on how to proceed in the non-prime dimensional case. 

I will also depict the relationship between Spekkens’ model, stabiliser quantum mechanics and Gross' theory of discrete Wigner functions (they are equivalent theories in odd dimensions) in terms of measurement update rules.

I will conclude by briefly discussing a project we have been recently working on that consists of characterising sub theories of Spekkens’ model that are operationally equivalent to sub theories of QM (in particular in the case of qubits) and use them to represent the non-contextual classically simulable part of state-injection schemes of computation with contextuality as a resource.

What is the essence of quantum theory? In the present talk I want to approach this question from a particular operationalist perspective. I take advantage of a recent convergence between operational approaches to quantum theory and axiomatic approaches to quantum field theory. Removing anything special to particular physical models, including underlying notions of space and (crucially) time, what remains is what I shall call "abstract quantum theory". This embeds into a more general framework that also includes classical theory, with classical and quantum theory representing two extremes in a spectrum of possible theories. I shall present this within a "hierarchy of abstraction", where adding structure leads to more specialized frameworks. In particular, adding topological spacetime and locality leads to the recently proposed positive formalism and recovers quantum field theory. Further rigidifying time and adding causality recovers the standard formulation of quantum theory. There are other promising specializations such as Hardy's proposal to work in op-space.

What can we say about the spectra of a collection of microscopic variables when only their coarse-grained sums are experimentally accessible? In this paper, using the tools and methodology from the study of quantum nonlocality, we develop a mathematical theory of the macroscopic fluctuations generated by ensembles of independent microscopic discrete systems. We provide algorithms to decide which multivariate gaussian distributions can be approximated by sums of finitely-valued random vectors. We study non-trivial cases where the microscopic variables have an unbounded range, as well as asymptotic scenarios with infinitely many macroscopic variables. From a foundational point of view, our results imply that bipartite gaussian states of light cannot be understood as beams of independent d-dimensional particle pairs. It is also shown that the classical description of certain macroscopic optical experiments, as opposed to the quantum one, requires variables with infinite cardinality spectra.

Computational complexity theory is a branch of computer science dedicated to classifying computational problems in terms of their difficulty. While computability theory tells us what we can compute in principle, complexity theory informs us regarding what is feasible. In this chapter I argue that the science of quantum computing illuminates the foundations of complexity theory by emphasising that its fundamental concepts are not model-independent. However this does not, as some have suggested, force us to radically revise the foundations of the theory. For model-independence never has been essential to those foundations. The fundamental aim of complexity theory is to describe what is achievable in practice under various models of computation for our various practical purposes. Reflecting on quantum computing illuminates complexity theory by reminding us of this, too often under-emphasised, fact.

Entropy is an important information measure. A complete understanding of entropy flow will have applications in quantum thermodynamics and beyond; for example it may help to identify the sources of fidelity loss in quantum communications and methods to prevent or control them. Being nonlinear in density matrix, its evaluation for quantum systems requires simultaneous evolution of more-than-one density matrix. Recently in [1] a formalism for such an evolution has been proposed and [2] shows that the flow of entropy between two systems corresponds to the full counting statistics of physical quantities that are exchanged between them. Interestingly, in quantum systems with heat dissipations this will not be equivalent to the second law of thermodynamics. In this talk I will describe a consistent formalism to evaluate entropy and show how to measure it in some quantum systems; for example in quantum point contacts in nanoelectronics and in the quantum heat engines introduced to describe photosynthesis and photovoltaic cells. The entropy flow is made of two parts: 1) an incoherent part, which can be re-evaluated semiclassically from the second law, and 2) a coherent part, which has no semiclassical analogue and appear as a result of extending Kubo-Martin-Schwinger (KMS) correlations.

[1] M.H.A. and Y. Nazarov, Phys. Rev. B 91, 104303 (2015)

[2] M.H.A. and Y. Nazarov, Phys. Rev. B 91, 174307 (2015)

When utilized appropriately, the path-integral offers an alternative to the ordinary quantum formalism of state-vectors, selfadjoint operators, and external observers -- an alternative that seems closer to the underlying reality and more in tune with quantum gravity. The basic dynamical relationships are then expressed, not by a propagator, but by the quantum measure, a set-function $\mu$ that assigns to every (suitably regular) set $E$ of histories its generalized measure $\mu(E)$. (The idea is that $\mu$ is to quantum mechanics what Wiener-measure is to Brownian motion.) Except in special cases, $\mu(E)$ cannot be interpreted as a probability, as it is neither additive nor bounded above by unity. Nor, in general, can it be interpreted as the expectation value of a projection operator (or POVM). Nevertheless, I will describe how one can ascertain $\mu(E)$ experimentally for any specified $E$, by means of an arrangement which, in a well-defined sense, acts as an $E$-pass filter. This raises the question whether in certain circumstances we can claim to know that the event $E$ actually did occur.

REFERENCE:

Alvaro Mozota Frauca and Rafael Dolnick Sorkin, How to Measure the Quantum Measure, Int J Theor Phys 56: 232-258 (2017), arxiv:1610.02087

In this talk I will discuss how we might go about about performing a Bell experiment in which humans are used to decide the settings at each end.  The radical possibility we wish to investigate is that, when humans are used to decide the settings (rather than various types of random number generators), we might then expect to see a violation of Quantum Theory in agreement with the relevant Bell inequality.  Such a result, while very unlikely, would be tremendously significant for our understanding of the world (and I will discuss some interpretations).  

Possible radical implications aside, performing an experiment like this would push the development of new technologies.  The biggest problem would be to get sufficiently high rates wherein there has been a human induced switch at each end before a signal as to the new value of the setting could be communicated to the other end and, at the same time, a photon pair is detected.   It looks like an experiment like this, while challenging, is just about feasible with current technologies.  

The fact that in physics concepts such as space, time, mass and energy are considered to be foundational has been conveniently serving a set of higher-level physical theories.

However, this keeps us from gaining a deeper understanding of such concepts which can in turn help us build a theory based on truly foundational concepts.

In this talk I introduce an alternate description of physical reality based on a simple foundational concept that there exist things that influence one another.

A network of objects that influence one another form a partially-ordered set (poset) that is called the influence network is considered. By consistently quantifying such a network with respect to a distinguished chain of events that is called an embedded observer, I demonstrate in relevant special cases that influence events can only be quantified by the familiar mathematics of space-time (Minkowski metric and Lorentz transformations), influence gives rise to basic concepts in Euclidean geometry such as direction, dimension and subspaces as well as the Pythagorean theorem, the dot product and geometrical figures. Thus a discrete version of some of the Euclidean geometry’s fundamental concepts is derived in this picture. Finally I talk about the concept of influence in quantum mechanics and how the case of a free particle is identical to the Feynman checkerboard problem for the electron which is known to give rise to the Dirac equation.

The study of thermodynamics in the quantum regime has in recent years experienced somewhat of a renaissance in our community. This excitement is fueled both by the fundamental nature of the subject as well as the potential for heat machines designed with quantum advantages. Here, I will suggest the study of quantum thermodynamics restricted to a Gaussian regime, with two primary goals in mind. The first is to understand just how restrictive Gaussian states and operations are in a thermodynamic context; as such operations are often the most easily implemented in the laboratory it is pertinent to understand what limitations they impose on us. The second is to use the computational simplicity of Gaussian quantum mechanics to examine thermodynamics under less ideal but more realistic assumptions, such as with finite bath sizes and time scales. We will present two works in this seminar, one focusing on each of these motivations.

Hayden and Van Dam showed that starting with a separable state in Alice and Bob’s state space and a shared entangled state in a common bipartite resource space, then using local unitary operations, it is possible to produce an entangled pair in the state space while at the same time only perturbing the shared entangled state by a small amount, which can be made arbitrarily small as the dimension of the resource space grows. They referred to this as “embezzling entanglement” since numerically it “appears" that the resource state was returned exactly.

It is natural to wonder if using an infinite dimensional resource space and local operations, one can return the resource state exactly while producing an entangled state in their state space. Whether or not you can achieve this phenomenon of “perfect embezzlement of an entangled state” depends on which mathematical model one uses to describe “local”.

We prove that perfect embezzlement is impossible in the tensor model but is possible in the commuting model. We then relate this to current work on the conjectures of Connes and Tsirelson about different models for quantum conditional probabilities.

This talk is based on joint work with R. Cleve, L. Liu and S. Harris.