First Results Of Quantum Natural Language Processing (NLP) Initiative With Intel
Results of an intense 14-month-long Quantum project between ICHEC and Intel Corporation have been published in the Institute of Physics Machine Learning: Science and Technology journal. The research was carried out by computational scientists Dr. Lee James O'Riordan and Myles Doyle who have demonstrate the use of combining classical and quantum computing models to perform Natural Language Processing tasks.
Natural Language Processing represents the automatic handling of natural (i.e. human) speech or text using AI. NLP is the technology behind applications which read and interpret verbal or written language driving applications such as Apple ‘SIRI’ and Amazon’s ‘Alexa’. NLP is used across sectors as diverse as healthcare, media, finance and human resources, and is an active area of research in AI. Advances in next-generation computing capabilities, such as quantum computing, will allow the exploration of more complex problems that are currently intractable even on current supercomputers.
Commenting on the results Dr. O’Riordan, the quantum activity lead at ICHEC, said, “Understanding how we adapt real-world problems to work on quantum computing devices is a hugely important, and often challenging, task. Developing machine learning methods using both classical and quantum computing models enables us to explore the strengths and weaknesses of these systems. For this work, we tasked ourselves to compare sentence meanings using such a hybrid classical-quantum method. For a chosen body of text, we demonstrated the similarity of the sentences based on the well-defined methods of distributional-compositional semantics.”
Commenting on the specifics of the research, computational scientist, Myles Doyle said:
“Using this distributional-compositional approach, we develop tools to use hybrid quantum techniques for comparing the meanings of tensor-composed sentence structures. With this we can represent small and large scale textual data sets to be encoded, processed, and decoded using a quantum circuit model. In addition, we release our developed toolkit as an open software suite to be used by the wider community, which is available on GitHub.”
Dr Venkatesh Kannan, Centre Technical Manager at ICHEC said:
“This work further develops an innovation environment in Ireland between industry and research organisations to develop expertise to program quantum computers. Intel and ICHEC identified the opportunity to use the Intel Quantum Simulator for projects like this and we believe there are more opportunities in this area we can jointly explore."
ICHEC is best positioned to foster these relationships as a founder of the Quantum Programming Initiative in Ireland. As part of this initiative ICHEC is involved in several additional Quantum projects”, he concluded.
Over the next 4 years, ICHEC will work in a consortium of European research and industry partners to further advance Quantum Natural Language Processing, alongside Quantum Chemistry for efficient absorption and mitigation of CO2 in the atmosphere. More on this exciting European Quantum Flagship project will be announced shortly.”