Quantum Programming Initiative

Research and development of quantum technologies has gained significant attention and momentum globally. The EU Quantum Flagship programme invests €1 billion over a 10 year period starting 2018, the US passed a bill in 2018 to invest USD 1.2 billion, a USD 270 million budget in Japan for a 10-year programme, and USD 10 billion commitment by China. Similarly sized commitments have been made by other countries including Australia, Germany, India, Singapore and the UK.

While the R&D activities in quantum technologies are still nascent globally (both in the academic and enterprise sectors), the work to develop the software ecosystem for quantum computing is the one that is least addressed. This has become apparent from discussions with the policy makers and directors in the European Commission’s Quantum Technologies programme, the QuantERA network, and other European academic and enterprise organisations in this space.

 

Strategy

 

In 2018, ICHEC identified the opportunity to be co-developers and early exploiters of the software systems for programming quantum computing platforms. To achieve this, ICHEC launched the Quantum Programming Initiative in 2018 with three primary lines of action:

  1. Development of practical applications on currently available quantum computing platforms (hardware and software simulators).
  2. R&D in software development to enable programming quantum computing platforms.
  3. Education/training to the academic and enterprise communities.

With the immediate availability of software quantum simulators and small-scale quantum hardware testbeds, it is essential that we develop the software ecosystem and programming expertise to target quantum platforms. It is crucial to do this simultaneously with ongoing efforts to develop larger-scale, reliable and commercially deployable quantum computers in the next 5-10 years.

As of June 2019, ICHEC has two multi-year collaborative projects with Intel® Corporation and Accenture Labs to research and develop machine learning and chemistry applications on existing quantum computing platforms. The projects are co-funded by the enterprises and the Irish state (through the Enterprise Ireland Industry Partnership Programme). Furthermore, ICHEC has designed hands-on quantum programming introductory workshops that can be offered over half, one or two days to academic and enterprise organisations.

Quantum Natural Language Processing (QNLP)

In 2018, Intel and ICHEC identified the opportunity to leverage the Intel Quantum Simulator to port and implement a quantum version of an existing Natural Language Processing (NLP) algorithm to analyse the meaning sentences in a corpus.

In this project, we implement two quantum algorithms on the Intel Quantum Simulator to address the “closest vector” and the “CSC sentence similarity” problems required in the quantum version of the DisCo model for NLP. The distributional compositional semantics (DisCo) model was originally developed by its authors with direct inspiration from quantum theory. The authors of DisCo have developed a quantum version of the algorithm that can be implemented on quantum computers.

Given a corpus, the implemented solution computes the meanings of two sentences (built from words in the corpus) and decides if their meanings match. We target corpora with ~2000 most common words using the Intel Quantum Simulator installed on the Irish national supercomputer (Kay) simulating up to ~40 qubits.

 

Meeting between Anne Matsuura (Intel) and Venkatesh Kannan (ICHEC) at Intel HQ (Santa Clara), April 2019
Anne Matsuura (Intel) and Venkatesh Kannan (ICHEC) at Intel HQ (Santa Clara), April 2019
Picture courtesy: JC Desplat (ICHEC)

More details at Quantum Natural Language Processing.

    Qauntum PFAS Chemicals Remediation (QPFAS)

    The PFAS are a toxic family of chemicals that enter the human consumption ecosystem through waterproof coatings, non-stick surfaces in surfaces/cookware, food packaging and Teflon coatings, and also the biological system through inhalation, ingestion and absorption via the skin. They are highly persistent and non-biodegradable. The health impacts include cancer (of kidney, bladder and liver), thyroid diseases, hormonal changes, increased cholesterol among other conditions. Therefore, analysis of remediation of PFAS from the ecosystem is crucial and widely studied through wet-lab experiments as well as computational simulation.

    This project with Accenture Labs focuses on the molecular simulation and electronic/chemical properties analysis of PFAS for their destruction through chemical oxidation/reduction. Currently, computational models and tools are used with classical HPC for simulation and analysis of molecular structures and their electronic/chemical properties. However, such classical HPC models are computationally expensive and not particularly scalable, which limits the accuracy and the size and number of compounds subject to study. Approximations need to be done to reduce the computation cost and also to map molecular and electronic properties, which are inherently quantum, for execution on classical computation platforms.

    This project aims at:

    1. Documenting the research of different methods to map the PFAS chemicals for their simulation on quantum computing platforms.
    2. Implementing a hybrid classical-quantum software solution for the molecular simulation of the PFAS family of chemicals (for their remediation) on currently available quantum computing platforms using existing software toolkits.
    3. Testing and evaluating the mapping methodologies and the software implementation using currently available quantum computing platforms.

    More details at Quantum PFAS Chemicals Remediation.

    Exascaling Intel-QS (ExIQS)

    The Intel Quantum Simulator (Intel-QS) is a distributed high-performance implementation of a quantum simulator on a classical computer. The original version of Intel-QS can simulate single-qubit gates and two-qubit controlled gates, and is composed of a number of single- and multi-node optimisations, including vectorization, multi-threading, cache blocking, as well as overlapping computation with communication.

    In this project, ICHEC works with Intel to extend the implementation of Intel-QS by addressing the following aspects:

    1. Single-node and multi-node scalability (strong and weak scaling) beyond the limits imposed by standard MPI implementations.
    2. Efficient implementation of additional higher-order gates.
    3. Simulator components for noise and error modelling.

    More details at Exascaling Intel-QS.

    Training in Quantum Programming

    ICHEC has also designed workshops to provide introduction to quantum computing and training for academic and enterprise organisations for quantum programming on multiple quantum computing platforms (Intel, Rigetti and IBM) using a variety of use-case applications.

    The workshop can be customised for delivery over a period of half, one or two days to audiences with beginner or intermediate experience in quantum computing.

    More details at Training in Quantum Programming.

    Press

    • "Forbidden light wave tool could lead to crazy-fast supercomputers", Silicon Republic, July 2019. [here]
    • "Within a Quantum", TechCentral.ie, June 2019. [here]
    • "Intel to support the Irish Centre for High End Computing on new collaborative quantum computing project", Intel Newsroom, March 2019. [here]
    • "Intel and Irish supercomputer group aim for a first in quantum computing", Silicon Republic, March 2019. [here]
    • "ICHEC gets quantum computing development funding", TechCentral, March 2019. [here]
    • "Ireland and Intel to Study Natural Language Processing Between Man and Machine Using Quantum Computing", The Qubit Report, March 2019. [here]
    • "Intel supports new quantum computing NLP research project in Ireland", Telecompaper, March 2019. [here]
    • "Ireland Focuses on Programming Practical Use-case Quantum Technologies", The Qubit Report, January 2019. [here]