lambeq (named after mathematician Joachim Lambek) is the result of research that has gone on for 10 years by Cambridge Quantum’s (CQ) Chief Scientist Bob Coecke. The concept is that a sentence consisting of a group of words can be represented as a network and the network can be converted to a quantum circuit in an application called Quantum Natural Language Processing (QNLP). This approach can enable real-world QNLP applications, such as automated dialogue, text mining, language translation, text-to-speech, language generation and bioinformatics. It can also be used for a more general analysis of symbol sequences that can be found in genomics and proteomics.

The lambeq software is an open sourced modular software library that can perform the transformation through a number of stages as shown in a diagram of the pipeline below. The lambeq software would implement the first four stages shown in gray in the diagram below. The last optimization step shown in orange would be performed by Cambridge Quantum’s tket software for a quantum computer or quantum simulator or to a machine learning library such as Pytorch or Jax for a classical computer.

Diagram of the lambeq Pipeline. Credit: Cambridge Quantum

To learn more about lambeq and the QNLP research being undertaken at Cambridge Quantum, you can view a news release announcing the release of lambeq here, two Medium articles that describe Cambridge Quantum QNLP technology here and here, an arXiv preprint that describes lambeq in more detail here and for those who want to try out the code, the lambeq GitHub library available here.

October 14, 2021