As we recently discussed in our ongoing look at linguistic AI, natural language processing (NLP) is an established technology with astonishing real-world applications in areas such as machine translation that nevertheless seems often to be on the sidelines of newsworthy developments in more experimental areas like natural language generation (NLG). Now, however, NLP is getting a new infusion from the cutting-edge of computer science with reports from Cambridge Quantum on their development of the first quantum natural language processing (QNLP) toolkit. Specifically, the software uses QNLP to turn sentences into quantum circuits, furthering the development of functions including text mining, automated dialogue, language translation, text-to-speech, bioinformatics, and language generation. Although this new and innovative technology is still early in steps of development, transitioning from the NLP using conventional computers to NLP using quantum software represents a milestone for anyone concerned with language AI technology.
As a discipline, quantum computing involves far more complex principles than traditional computing, such as the use of qubits rather than the traditional binary 0s and 1s and unconventional interpretations of the basic laws of physics. Even more, probabilities and uncertainty are building blocks of the software’s inner workings, enabling the QNLP to address many of the current challenges in equally experimental fields like NLG. Fittingly named after the late Jaochim Lambek, a notable professor of computer science, math, and linguistics at McGill University, this QNLP toolkit dubbed “lambeq” further adds a new perspective and computing abilities to the basic functions of one of the most incorporated linguistic algorithms in the translation industry.
As we covered in the past, NLP is ingrained in many mainstream translation services, such as Google Translate, and functions as the basic tool offered by many language service providers (LSPs). Although the work by Cambridge Quantum is just the latest in technological advancements of linguistic AI, the general purpose of using these advanced systems as a translation tool is still relatively new and must keep up with the changing ways of language. As with other challenges in evolving language AI, being able to detect and process culturally unique language patterns and applying these to contemporary translation platforms is still an indication that there is a long way until human translators will play a lesser role in translation services. Despite this, from a LSP perspective, QNLP is groundbreaking in that it takes an already well-established language processing technology and pushes it beyond the boundaries and capabilities of a traditional computer. AI’s growing role in global communications and localization services continues to change in ways unseen in the past and offers new opportunities for LSPs to target specialized markets. Ultimately, advancing state of the art language AI is likely to continue with this trend, as companies like CSOFT monitor the developments of today that could change the world tomorrow.
To learn more about CSOFT’s translation services and technology-driven localization solutions, visit www.csoftintl.com!