An effective, accessible abstraction hierarchy has made using and programming classical computers possible for people across all disciplines. It is our hope that the same can be achieved for quantum computing. However, current quantum programming languages and frameworks often require users to have PhD-level knowledge of information theory, mathematics, physics, and/or chemistry.
At the applications level, the prevalence of gate-based reasoning to develop quantum algorithms hampers non-experts in delving into the domain.
At the hardware level, frameworks are tied closely to the operations each physical device happens to implement. To push the field of quantum computing forward, we deem it essential for the community to develop a common framework and hierarchy of abstractions for use by platform developers and quantum programmers.
The workshop seeks to bring together experts spanning the entire quantum software stack with the aim of tackling this challenge.
We highlight the necessity for:
a more user-friendly and information-rich programming language,
a hardware abstraction that decouples from restrictions of physical implementations
an execution model tying the two levels together.
The workshop will have three sessions that reflect the organizers’ current, high-level picture of the quantum abstraction hierarchy. Specifically, the session topics focus on understanding the perspectives of users and how they see the field from their position in this hierarchy:
Top-down: What do developers of applications and programming frameworks need and expect from lower levels of the stack?
Bottom-up: What do hardware providers and low-level compiler designers expect from higher levels of the stack?
Meet in the middle: Can we identify and prioritize the important problems and facilitate productive interactions at the hardware-software interface?
Here is a list of important topics to be addressed:
Beyond gate-level abstractions: What are the limitations of current thinking in quantum gates and matrices that are hampering the development of quantum workflows? Could new abstractions spark more creativity? Are the programming models we have today going to be enough for the domain to thrive, or are they limiting future innovation and mainstream adoption?
Seamless quantum–classical integration: Can we imagine an integration pattern different than the accelerator/co-processor role for the quantum device, and what kinds of adjustments in programming and system organization would that entail?
Reconciling different perspectives: Is there a way we can have software engineers, hardware engineers, and physicists agree on a single, unifying programming model?
Classical paradigms vs. quantum thinking: Has the way we have structured quantum programming to resemble classical programming really simplified the process for newcomers, or does it limit the expressiveness of quantum algorithms?
Abstraction vs. performance trade-off: Does abstracting away hardware characteristics limit the performance of device-specific optimizations? What could be an acceptable trade-off?
Premature commitment and optimization: How can we agree on a set of conventions that do not bind us to a standard but rather enable productive collaboration? Is it too early to agree on these details? Is this a problem that will solve itself if left alone?
Accessibility vs. complexity: Should we simplify quantum algorithm development, as has been done for classical computing, or is the domain inherently so complex that any abstraction would hamper essential understanding?
Olivia Di Matteo, University of British Columbia
Edo Giusto, University of Naples Federico II
Michał Stechły, PsiQuantum
Scott Pakin, Los Alamos National Laboratory
Santiago Núñez-Corrales, NCSA/UIUC
Vlad Știrbu, University of Jyväskylä