Quantum theory from information inference principles

Tuesday, Nov 11th
Philipp Hoehn, Perimeter Institute 
Title: Quantum theory from information inference principles 
PDF of the talk (800k)
Audio [.wav 40MB]

by Matteo Smerlak, Perimeter Institute

When a new theory enters the scene of physics, a succession of events normally takes place: at first, nobody cares; then a minority starts playing with the maths while the majority insists that the theory is obviously wrong; farther down the road, we find the majority using the maths on a daily basis and all arguing that the theory is so beautiful, it can only be right; along the way, thanks to many years of practice, a new kind of intuition grows out of the formalism, and our entire picture of reality changes accordingly. This is the process of science.

For some reason, though, the eventual shift from formalism to intuition never happened for quantum mechanics (QM). Ninety years after its discovery, specialists still call QM “weird”, teachers still quote Feynman claiming that “nobody really understands QM”, and philosophers still discuss whether QM requires us to be “antirealist”, “neo-Kantian”, “Bayesian”… you name it. Niels Bohr wanted new theories to be “crazy enough”, but it seems this one is just too crazy. And yet it works!

In the face of this puzzle, a school of thought initiated by Birkhoff and von Neumann in the thirties has declared it its mission to reconstruct QM. The idea is simple: if you don’t get how the machine works, then roll up your sleeves, take the machine apart, and build it again—from scratch. Indeed this is how Einstein delt with the symmetry group of Maxwell’s equations (and its mysterious action on lengths and durations): he found intuitive two physical principles—the relativity principles—and derived the Lorentz group (the set of symmetries of Maxwell's equations) from them. Thus special relativity was “really understood”.

Much recent work towards a reconstruction of QM has taken place within a framework called “generalized probability theories” (GPT). This approach elaborates on basic notions such as preparations, transformations and measurements. The main achievement of GPT has been to locate QM within a more general landscape of possible modifications of classical probability theory. It has showed for instance that QM is not the most non-local theory consistent with what is known as no-signaling property: stronger correlations than quantum entanglement are in principle possible, though they are not realized in nature. To understand what is, we must know what else could have been—thus speak GPT proponents. 

Philipp uses a different language for his reconstruction of QM: instead of measurements and states, he talks about questions and answers. The semantic shift is not innocent: while a “measurement” uncovers the intrinsic state of a system, a “question” only brings information to whoever asks it—that is, a question relates to two entities (the system and the observer/interrogator) rather than just one (the system). Because there isn’t anybody out there to ask questions about everything, there is no such thing as the “state of the universe”, Philipp says!

This so-called “relational” questions/answers approach to QM was advocated twenty years ago by Rovelli, who emphasized its similarity with the structure of gravitation (time is relative, remember?). He also proposed two basic informational principles: one states that the total information that an observer O can gather about a system S is limited; the second specifies that, even when O has obtained the maximum amount of information about S, she can still learn something about S by asking other, “complementary” questions. Thence non commuting operators! Similar ideas where discussed independently by Zeilinger and Brukner—and Philipp embraces them wholeheartedly.

But he also takes a big step further. Adding four more postulates to Rovelli’s (which he calls completeness, preservation, time evolution and locality), Philipp shows how to reconstruct the set Σ of all possible states of S relative to O (together with its isometry group, representing possible time evolutions). For a quantum system allowing only one independent question—a qubit—Σ is a three-dimensional ball, the Bloch sphere. (Note that a 3-ball is a much bigger space than a 1-ball, the state space of a classical bit—enter quantum computing…) For systems with more independent questions, i.e. N qubits, Σ is the mathematical structure known as the convex cone over some complex projective space—not quite what is known as a Calabi-Yau manifold, but still a challenge for the mind to picture.

N=2 turns out to be the most difficult case: once this one is solved—Philipp says this took him a full year, with inputs from his collaborator Chris Wever—, higher N’s follow rather straightforwardly. This is a reflection of a crucial aspect of QM: quantum systems are “monogamous”, meaning that they can establish strong correlations (aka “entanglement”) with just one partner at a time. Philipp’s questions/answers formulation provides a new and detailed understanding of this peculiar correlation structure, which he represents as a spherical tiling. “QM is beautiful!”, says Philipp.

One limitation of Philipp’s current approach—also pointed out by the audience—is the restriction to binary (or yes/no) questions. A spin-1 particle, for instance, falls outside this framework, for it can give three different answers to the question “what is your spin in the z direction?”, namely “up”, “down” or “zero”. Can Philipp deal with such ternary question, and reconstruct the 8 dimensional state space of a quantum “trit”? We wish him to find the answer within… less than a year! 

Holographic special relativity: observer dependent geometry

Derek Wise, FAU Erlangen
Title: Holographic special relativity: observer space from conformal geometry 
PDF of the talk (600k) Audio [.wav 38MB] Audio [.aif 4MB] 

by Sean Gryb, Radboud University

Introduction

In Roman mythology, Janus was the god of gateways, transitions, and time, whose two distinct faces are depicted peering in opposite directions, as if bridging two different regions (or epochs) of the Universe. The term “Janus-faced” has come to mean a person or thing that simultaneously embodies two polarized features, and the Janus head has come to represent the embodiment of these two distinct features into one.

In this talk (based off the paper [1]), Derek Wise explores the possibility that spacetime itself might be Janus-faced. He explores an intriguing relationship between the structure of expanding spacetime and the scale-invariant description of a sphere. What he finds is a mathematical relationship providing a bridge between these two Janus faces that distinctly represent events in the Universe. This bridge is remarkably similar to the picture of reality proposed by the holographic principle and, in particular, the AdS/CFT correspondence where, on one side, there is the usual spacetime description of events and, on the other, there is a way to imprint these events onto the 3-dimensional boundary of this spacetime.

Aside from providing an alternative to spacetime, Derek's picture may even help illuminate the deeper structures behind a recent formulation of general relativity (GR) called Shape Dynamics, which I will come to at the end of this post. But to begin, I will try to explain Derek's result by first giving a description of the spacetime aspect of the Janus face and then describe how a link can be established to a completely distinct face, which, as we will see, is a description of events in the Universe that is completely free of any notion of scale. The key points of the discussion are summarized beautifully in the depiction of Janus given below, by Marc Ngui, who has provided all the images for this post. The diagram shows how, as I will describe later, events seen by observers in spacetime can be described by information on the boundary. I encourage the reader to revisit this image as its main elements are progressively explained throughout the text.

Relativity, Observers, and Spacetime

In 1908, Hermann Minkowski made a great discovery: Einstein's new theory of Special Relativity could be cast into a beautiful framework, one that Minkowski recognized as a kind of union of space and time. In his own words: “space by itself, and time by itself, are doomed to fade away into mere shadows, and only a kind of union of the two will preserve an independent reality.”[2] To understand what Minkowski meant, let's go back to 1904-1905 in order to retrace the discoveries that spawned Minkowski's revolution.

Relativity concerns the way in which different observers organize information about ‘when’ and ‘where’ events take place. Einstein realized that this system of organization should have two properties: i) it should work the same way for each observer, and ii) it should involve a set of rules that allows different observers to consistently compare information about the same events. This means that different observers don't necessary need to agree on when and where a particular event took place, but they do need to agree on how to compare the information gathered by different observers. Einstein expressed this requirement in his principle of relativity, to which he gave primary importance within physical theories. The key point is that relativity is fundamentally a statement about observers and how they collect and compare information about events. Minkowski's conception of spacetime comes afterwards, and it comes about through the specific mathematical properties of the rules used to collect and compare the relevant information.

To try to understand how spacetime works, we will use a slightly more modern version of spacetime than the one used by Minkowski — one with all the same essential properties as the original, but which can accommodate the observed accelerated expansion of space. This kind of spacetime was first studied by Willem de Sitter, and is named de Sitter (dS) spacetime after him. It has the basic shape depicted by the blue grid in the Janus image above. Because this space is curved, it is most convenient to describe it by putting it into a larger dimensional flat space (just like the 2D surface of a sphere depicted in a 3D space). This means that we can label events in this spacetime by 5 numbers: 4 space components, labeled (x, y, z, w) and one time component, t, that obey the relation

x² + y² + z² + w² - t² = ℓ².    (1)
This restriction (which serves as the definition of this spacetime) means that the 4 space components are not all independent. Indeed, the single constraint above removes one independent component, leaving the 3 space directions we know and love. The parameter ℓ is related to the cosmological constant and dictates how fast space is expanding. Adjusting its value changes the shape of the spacetime as illustrated in the figure below.

The middle spacetime in blue depicts a typical dS spacetime. Increasing the parameter ℓ decreases the rate of expansion so that, if ℓ → ∞, the spacetime barely expands at all and looks more like the purple cylinder on the right. The opposite extreme, when ℓ → 0, is the yellow light cone, which is named that way because the space is expanding at its maximum rate: the speed of light. This extreme limit will be very important for our considerations later.

Although this model of spacetime is dramatically simplified, it remarkably describes, to a good approximation, two important phases of our real Universe: i) the period of exponential expansion (or inflation), which we believe took place in the early history of our Universe, and ii) the present and foreseeable future. Different observers compare the labels they attribute to each event by performing transformations that leave the form of (1), and thus the shape of dS spacetime, unchanged. Because of this property, these transformations constitute symmetries of dS spacetime. Since the transformations that real observers must use to compare information about real events just happen to correspond to spacetime symmetries, it is no wonder that the notion of spacetime has had such a profound influence on physicists’ view of reality. However, we will shortly see that these rules can be recast into a completely different form, which tells a different story of what is happening.

Spacetime's Janus Face

We will now see how the symmetries of observers in dS spacetime can be rewritten in terms of symmetries that preserve angles, but not necessary distances, in space. In particular, all information about scale is removed. In mathematics, these are called conformal symmetries. This means that different observers have a choice when analyzing information that they collect about events: either they can imagine that these events have taken place in dS spacetime, and are consequently related by the dS symmetries; or they can imagine that these events are representing information that can be expressed in terms of angles (and not lengths), and are consequently related by conformal symmetries.

To understand how this can be so, consider the very distant future and the very distant past: where dS spacetime and the light cone nearly meet. This extreme region is called the conformal sphere because it is a sphere and also because it is where the dS symmetries correspond to conformal symmetries.

In fact, any cross-section of the light cone formed by cutting it with a spatial plane (as illustrated in the diagram below) is a different representative of the conformal sphere since these different cross-sections will disagree on distances but will agree on angles. Although the intersection looks like a circle (represented in dark green), it is actually a 3-dimensional sphere because we have cut out 2 of the spatial dimensions (which we can't draw on a 2 dimensional page).

To see how events on this 3d sphere can be represented in a scale-invariant way on a 3d plane, we can use a handy technique called a stereographic projection. The stereographic projection is often used for map drawing where the round earth has to be drawn onto a flat map. One of its key properties, namely that it preserves angles, means that maps drawn in this way are useful for navigating since an angle on the map corresponds to the same angle on the Earth. It is precisely this property that will make the stereographic projection useful for us here.

To perform a stereographic projection, imagine picking a point on a sphere, which we can interpret as the location of a particular observer on the sphere (represented by an eye in the diagram below), and call this the South Pole. Now imagine putting a light on the North Pole and letting it shine through the space that the sphere has been drawn in. Suppose our sphere is filled with points. Then, the shadow of these points will form an image on the plane tangentially to the sphere on the South Pole. The picture below illustrates what is going on. Points on the sphere are represented by stars and the yellow rays indicate how their image is formed on the plane.

It is now a relatively straightforward mathematical exercise to show that the symmetries of the light cone represent transformations on the plane that may change the size of the image, but will preserve the angles between the points. Thus, the symmetries of the cone can be understood in terms of the conformal symmetries of this plane.

If we now move our cross-section ever further into the future or the past, then the dS spacetime begins to resemble more and more the light cone. Thus, if we can represent arbitrary events in dS spacetime by information imprinted on two cross-sections in the infinite future and infinite past, then these events can be represented in terms of the images they induce onto our projected planes, and we have obtained our objective.

There is a simple way that this can be done. Imagine taking, as shown in the figure below, an arbitrary event in dS spacetime and drawing all the events in the distant past that could affect things that happen at this point (this region is a finite portion of the spherical cross-sections because no disturbance can travel faster than the speed of light). The result is a 2 dimensional spherical region, called the particle horizon indicated by the red regions in the diagram below, which grows steadily over time. You can think of this region as the proportion of dS spacetime that is visible at any particular place. In fact, you can use the relative size of this region as an indication of the time at which that event occurs. Because this is a notion of time that exists solely in terms of quantities defined in the distant past, it will transform under conformal symmetries. To give an idea of what this looks like, the motion of an observer from some point in the distant past to a new point in the distant future is represented by a series of concentric spheres, starting at the initial point and then spreading out to eventually cover the whole sphere. The diagram below shows how this works. The different regions (a,b,c,d) represent progressively growing regions corresponding to progressively later times.

In this way, you can map information about events in dS spacetime to information on the conformal sphere. In other words, the picture of reality that one gets from Einstein's theory of Special Relativity is a story that can be told in two very different ways. In the first way, there are events which trace out histories in spacetime. ‘Where’ and ‘when’ a particular event takes place depends on who you are, and the information about these events can be transformed from one observer to another via the global symmetries of spacetime. In the new picture, it is the information about angles that is important. ‘Where’ and ‘how big’ things are depends on your point of view and the information about particular events can be transformed from one observer to another using conformal transformations.

From Special to General Relativity

We have just described how to relate two very different views of how observers can collect information about the world. Until now, we have only been considering homogeneous spaces: i.e., those that look the same everywhere. The class of observers we were able to consider was resultantly restrictive. It was Einstein's great insight to recognize that the same mathematical machinery needed to describe events seen by arbitrary observers could also be used to study the properties of gravity. The machinery in question is a generalization of Minkowski's geometry, named after Riemann.

In order to describe Riemannian geometry, it is easiest to first describe a generalization of it (which we will need later anyway), and then show how Riemannian geometry is just a special case. The generalization in question is called Cartan geometry, after the great mathematician Élie Cartan. Cartan had the idea of building general curved geometries by modelling them off homogeneous spaces. The more general spaces are constructed by moving these homogeneous spaces around in specific ways. The geometry itself is defined by the set of rules one needs to use to compare vectors after moving the homogeneous spaces. These rules split into two different kinds: those that change the point of contact between the homogeneous space and the general curved space and those that don't. These different moves are illustrated for the case where the homogeneous space is a 2D sphere in the diagram below.
 The moves that don't change the point of contact (in the case above, this corresponds to spinning about the point of contact without rolling) constitute the local symmetries of the geometry and could, for example, correspond to what different local observers would see (in this case, spinning observers versus stationary ones) when looking at objects in the geometry. Einstein exploited this kind of structure to implement his general principle of relativity described earlier. The moves that change the point of contact (in the case above, this means rolling the ball around without slipping) give you information about the curved geometry of the general space. Einstein used a special case of Cartan geometry, which is just Riemannian geometry, where the homogenous space is Minkowski space. He then exploited the analogue of the structure just described to explain an old phenomenon in a completely new way: gravity. In the process, he produced one of our most radical yet successful theories of physics: General Relativity. The figure below shows how the different kinds of geometry we've discussed are related.

Now, consider what happens when we substitute, as we did in the last section, Minkowski's flat spacetime for de Sitter's curved, but still homogenous, spacetime. We can still describe gravity, but in a way that naturally includes a cosmological constant. However, the conformal sphere is also a homogeneous space. Moreover, as we described earlier, the symmetries of this homogeneous space can be related to the dS symmetries. This suggests that it might be possible to describe gravity in terms of a Cartan geometry modelled off the conformal sphere.

From the Conformal Sphere to Shape Dynamics?

Cartan geometries modelled on the conformal sphere are called conformal geometries because the local symmetries of these geometries preserve angles, and not scale. Although we have laid out a procedure relating the model space of conformal geometries to the model space of spacetimes with a cosmological constant, it is quite another thing to rewrite gravity in terms of conformal geometry. This is, in part, because the laws governing spacetime geometry are complicated and, in part, because our prescription for relating the model spaces is also not straightforward, since it relates local quantities in spacetime to non-local quantities in the infinite future and past. Nevertheless, this exciting possibility provides an interesting future line of research. Furthermore, there are other hints that such a description might be possible.

Using very different methods, it is possible to show that General Relativity is actually dual to a theory of evolving conformal geometry [3]. However, the kind of conformal geometry used in this derivation has not yet been written in terms of Cartan geometry (which makes use of slightly different structures). This new way of describing gravity, called Shape Dynamics, is perhaps making use of the interesting relationship between spacetime symmetries and conformal symmetries described here. Understanding exactly the nature of the conformal geometry in Shape Dynamics and its relation to spacetime could prove valuable in being able to understand this new way of describing gravity. Perhaps it could even be a window into understanding how the quantum theory of gravity should work?

• [1] D. K. Wise, Holographic Special Relativity, arXiv:1305.3258 [hep-th].
• [2] H. Minkowski, The Principle of Relativity: A Collection of Original Memoirs on the Special and General Theory of Relativity, ch. Space and Time, pp. 75–91. New York: Dover, 1952.
• [3] H. Gomes, S. Gryb, and T. Koslowski, Einstein gravity as a 3D conformally invariant theory, Class. Quant. Grav. 28 (2011) 045005, arXiv:1010.2481 [gr-qc].

-->

Spectral dimension of quantum geometries

Johannes Thürigen, Albert Einstein Institute 
Title: Spectral dimension of quantum geometries 
PDF of the talk (1MB) Audio [.wav 39MB] Audio [.aif 4.1MB] 

By Francesco Caravelli, University College London

One of the fundamental goals of quantum gravity is understanding the structure of space-time at very short distances, together with predicting physical and observable effects of having a quantum geometry. This is not easy. Since the introduction of fractal dimension in Quantum Gravity, and its importance emphasized in the work done in Causal Dynamical Triangulations (Loll et al. 2005) and Asymptotic Safety (Lauscher et al. 2005), it has become more and more clear that space-time, at the quantum level, might have a radical transformation: the number of effective dimensions might change with the energy of the process involved. Various approaches to Quantum Gravity have collected evidences of a dimensional flow at high energies, which was popularized by Carlip as Spontaneous Dimensional Reduction (Carlip 2009, 2013). (The use of the term reduction is indeed a hint that a dimensional reduction is observed, but the evidences are far from conclusive. We find dimensional flow more appropriate.)

Before commenting on the results obtained by the authors of the paper discussed in the seminar
(Calcagni, Oriti, Thuerigen 2013), let us first step back for a second and spend some time introducing the concept of fractal dimension, which is relevant to this discussion.

The concept of non integer dimension was introduced by the mathematician Benoit Mandelbrot half a century ago. What is this fuss about fractals and complexity? What is the relation with spacetimes, quantum space-times?

Everything start from an apparently simple question asked by Mandelbrot: What is the length of the coast of England (or more precisely, Cornwall)? As it turned out, the length of the cost of England, depended on the lens used to magnify the map of coast, and depending on the magnifying power, the length changed with a well defined rule, known as scaling, which we will explain shortly.

There are several definitions of fractal dimension, but let us try to keep things as easy as possible, and see why a granular space-time might indeed imply a different dimensions at different scales (i.e., our magnifying power). The easy case is the one of a regular square lattice, which for the sake of clarity we consider infinite in any direction.

                                                          Source: Manny Lorenzo

 The dimension of the lattice might look two dimensional, as the lattice is planar: it can be embedded into a two dimensional surface (this is what is called embedding dimension). However, if we pick any point of this lattice, and count how many points are at a distance “d” from it, we will see that the number of points increases with a scaling law, given by*:

N ~ d^gamma .

If d is not too big, the value of gamma changes if the underlying structure is not a continuum, or is granular, and gamma can take non-integer values. This can be interpreted in various ways. For the case of fractals, this implies that the real dimension of fractals is not integer. Analogously to the case of the number of points within a certain distance d, it is possible to define a diffusion operation which will do the work of counting for us. However, the counting process depends on the operator which defines the diffusion process: how a swarm of particles move on the underlying discrete space. This is a crucial point of the procedure.

In the continuum, the technology is developed to the point that it can  to show that such an operator can be defined precisely**. The problem then is that the scaling not precise: for too long times, the scaling relation is not exact (as curvature effect might contribute). Thus, the time given to the particle to diffuse has to be appropriately tuned. This is what the authors define in Section 2 of the paper discussed in the talk and is a standard procedure in the context of the spectral dimension. Of course, what discussed insofar is valid for classical diffusion, but the operator can be defined for quantum diffusion as well, which is, put in simple terms, described by a Schroedinger unitary evolution like in ordinary quantum mechanics.

It is important to understand that the combinatorial description of a manifold (how these are represented in the discrete setting), rather than the actual geometry, plays a very relevant role. If you calculate the fractal dimension of these lattices, although at large scale they give the right fractal dimension, on small scale they do not. This shows that in fact discreteness does have an effect on the spectral dimension, and that results do indeed depend on the number of dimensions. But more importantly the authors observe that the spectral dimension, even in the classical case, depends on the precise structure of the underlying pseudo-manifold, i.e. how the manifold is discretized. If you combine this with the fact that insofar the fractal dimension is the global observable saying in how many dimensions you live in (concept very important for other high energy approaches), the interest might be quite well justified.

The case of a quantum geometry, considered using Loop Quantum Gravity (LQG), is then put forward at the end. The definition is different from the one given previously (Modesto 2009, assuming that the scaling is given by the area operator of LQG), and it leads to different results.

Without going into the details (described anyway quite clearly in the paper), probably it is noteworthy to anticipate the results and explain the difficulties involved in the calculation. The first complication comes from the calculation itself: it is in fact very hard to calculate the fractal dimension in the full quantum case. However, in the semiclassical approximation (when the geometry is in part quantum and in part classical), the main "quantum" part can be neglected. The next issue is that, in order to claim the emergence of a clear topological dimension, the fractal dimension has to be constant for a wide range of distances of several orders of magnitude. It is important to say that, if you use the fractal dimension as your definition of dimension, it is not possible to assign a given dimensionality unless the number of discrete points under consideration is large enough. This is a feature of the fractal dimension which is very important for Loop Quantum Gravity in many respects, as there as been for long time a discussion on what is the right description of classical and quantum spacetime. Still, this approach gives the possibility of a bottom-up definition of dimension (in the top-down, there would not be any dimensional flow).

As a closing remark, it is fair to say that this paper goes one step further into defining a notion of fractal dimension in Loop Quantum Gravity. The previous attempt was made by Modesto and collaborators using a rough approximation to the Laplacian. That approximation exhibited a dimensional flow towards an ultraviolet 2-dimensional space, which seems to be not present using a more elaborated Laplacian.

 *For a square lattice, if d is big enough, \gamma is equal to two: this is the Haussdorf dimension of the lattice, and indeed this dimension can be defined through the following equation: gamma=\partial log(N)/ \partial d

** Using the technical terminology, this is the Seeley-De Witt expansion of the heat kernel on curved manifolds. This is usually called spectral dimension. The first term of the expansion depends explicitly on the spectral dimension, while in the terms at higher orders there are also contributions from the curvature.