Source author record

Jacques Pienaar

Jacques Pienaar appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

quant-ph gr-qc Machine Learning Mathematical Software physics.chem-ph physics.hist-ph Programming Languages

Catalog footprint

What is connected

8works

7topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2021arXiv

COMET: A Domain-Specific Compilation of High-Performance Computational Chemistry

The computational power increases over the past decades havegreatly enhanced the ability to simulate chemical reactions andunderstand ever more complex transformations. Tensor contractions are the fundamental computational building block of these simulations. These simulations have often been tied to one platform and restricted in generality by the interface provided to the user. The expanding prevalence of accelerators and researcher demands necessitate a more general approach which is not tied to specific hardware or requires contortion of algorithms to specific hardware platforms. In this paper we present COMET, a domain-specific programming language and compiler infrastructure for tensor contractions targeting heterogeneous accelerators. We present a system of progressive lowering through multiple layers of abstraction and optimization that achieves up to 1.98X speedup for 30 tensor contractions commonly used in computational chemistry and beyond.

preprint2020arXiv

MLIR: A Compiler Infrastructure for the End of Moore's Law

This work presents MLIR, a novel approach to building reusable and extensible compiler infrastructure. MLIR aims to address software fragmentation, improve compilation for heterogeneous hardware, significantly reduce the cost of building domain specific compilers, and aid in connecting existing compilers together. MLIR facilitates the design and implementation of code generators, translators and optimizers at different levels of abstraction and also across application domains, hardware targets and execution environments. The contribution of this work includes (1) discussion of MLIR as a research artifact, built for extension and evolution, and identifying the challenges and opportunities posed by this novel design point in design, semantics, optimization specification, system, and engineering. (2) evaluation of MLIR as a generalized infrastructure that reduces the cost of building compilers-describing diverse use-cases to show research and educational opportunities for future programming languages, compilers, execution environments, and computer architecture. The paper also presents the rationale for MLIR, its original design principles, structures and semantics.

preprint2020arXiv

Quantum causal models via QBism

This paper presents a framework for Quantum causal modeling based on the interpretation of causality as a relation between an observer's probability assignments to hypothetical or counterfactual experiments. The framework is based on the principle of `causal sufficiency': that it should be possible to make inferences about interventions using only the probabilities from a single `reference experiment' plus causal structure in the form of a DAG. This leads to several interesting results: we find that quantum measurements deserve a special status distinct from interventions, and that a special rule is needed for making inferences about what would happen if they are not performed (`un-measurements'). One natural candidate for this rule is found to be an equation of importance to the QBist interpretation of quantum mechanics. We find that the causal structure of quantum systems must have a `layered' structure, and that the model can naturally be made symmetric under reversal of the causal arrows.

preprint2018arXiv

Comment on `The notion of locality in relational quantum mechanics'

A recent paper [P. Martin-Dussaud, C. Rovelli, F. Zalamea, arXiv:1806.08150] has given a lucid treatment of Bell's notion of local causality within the framework of the relational interpretation of quantum mechanics. However, the authors went on to conclude that the quantum violation of Bell's notion of local causality is no more surprising than a common cause. Here, I argue that this conclusion is unwarranted by the authors' own analysis. On the contrary, within the framework outlined by the authors, I argue that the implications of the relational interpretation are much more radical.

preprint2016arXiv

A relational approach to quantum reference frames for spins

In the literature on quantum reference frames, the internal (relative) properties of a system are defined as those which are preserved under an arbitrary change of reference frame. For a system of quantum spins, these are all properties preserved by proper spatial rotations of the laboratory. However, this approach does not account for the hypothetical possibility of the laboratory becoming entangled to the system, as described by a second laboratory (the `Wigner's friend' scenario), in which case the relationship between the two laboratories is not a rotation, but is fundamentally quantum. To overcome this limitation, we re-define the reference frame transformations to be those that preserve the fidelities between subsystems. This enables us to derive U(2) as the correct symmetry group for transformations of a system of N spin-half particles. Next, we propose that systems having the same internal properties should be regarded as physically equivalent in the absence of an external frame. Remarkably, this implies that a single spin in a superposition relative to a spin magnet is equivalent to a macroscopic superposition of the magnet relative to the spin. We discuss the implications of this result for the Wigner's friend paradox.

preprint2015arXiv

A separable, dynamically local ontological model of quantum mechanics

A model of reality is called separable if the state of a composite system is equal to the union of the states of its parts, located in different regions of space. Spekkens has argued that it is trivial to reproduce the predictions of quantum mechanics using a separable ontological model, provided one allows for arbitrary violations of `dynamical locality'. However, since dynamical locality is strictly weaker than local causality, this leaves open the question of whether an ontological model for quantum mechanics can be both separable and dynamically local. We answer this question in the affirmative, using an ontological model based on previous work by Deutsch and Hayden. Although the original formulation of the model avoids Bell's theorem by denying that measurements result in single, definite outcomes, we show that the model can alternatively be cast in the framework of ontological models, where Bell's theorem does apply. We find that the resulting model violates local causality, but satisfies both separability and dynamical locality, making it a candidate for the `most local' ontological model of quantum mechanics.

preprint2014arXiv

A graph-separation theorem for quantum causal models

A causal model is an abstract representation of a physical system as a directed acyclic graph (DAG), where the statistical dependencies are encoded using a graphical criterion called `d-separation'. Recent work by Wood & Spekkens shows that causal models cannot, in general, provide a faithful representation of quantum systems. Since d-separation encodes a form of Reichenbach's Common Cause Principle (RCCP), whose validity is questionable in quantum mechanics, we propose a generalised graph separation rule that does not assume the RCCP. We prove that the new rule faithfully captures the statistical dependencies between observables in a quantum network, encoded as a DAG, and is consistent with d-separation in a classical limit. We note that the resulting model is still unable to give a faithful representation of correlations stronger than quantum mechanics, such as the Popescu-Rorlich box.

preprint2013arXiv

Causality Violation and Nonlinear Quantum Mechanics

It is currently unknown whether the laws of physics permit time travel into the past. While general relativity indicates the theoretical possibility of causality violation, it is now widely accepted that a theory of quantum gravity must play an essential role in such cases. As a striking example, the logical paradoxes usually associated with causality violation can be resolved by quantum effects. We ask whether the explicit construction of a theory that allows causality violation might in turn teach us something about quantum gravity. Taking the toy model of Deutsch as a starting point, in Part I we argue that, despite being a nonlinear modification of quantum mechanics, the model does not imply superluminal signalling and its predictions can be operationally verified by experimenters within an appropriate ontological setting. In Part II we show that the model can be directly applied to scalar quantum fields, provided the fields are sent back in time by an amount much larger than the coherence time of the wavepackets. We propose a generalisation of the model that lifts this constraint and recovers the predictions of standard quantum field theory when the size of the temporal jump is much smaller than the coherence time of the wavepackets. Finally, we discuss an extension of the model to ordinary gravitational time dilation. The resulting model generalises and extends earlier work on the topic and can be tested experimentally using current technology.