Parts of Speech Tagging in NLP – Runtime Optimization with Quantum Formulation and ZX Calculus ARIT KUMAR BISHWAS, Amity Institute of Information Technology, Noida, India ASHISH MANI, Department of EEE, Amity University, Noida, India VASILE PALADE, Research Centre for Data Science, Coventry University, Coventry, UK This paper proposes an optimized formulation of the parts of speech tagging in Natural Language Processing with a quantum computing approach and further demonstrates the quantum gate-level runnable optimization with ZX-calculus, keeping the implementation target in the context of Noisy Intermediate Scale Quantum Systems (NISQ). Our quantum formulation exhibits quadratic speed up over the classical counterpart and further demonstrates the implementable optimization with the help of ZX calculus postulates. KEYWORDS Natural Language Processing, Quantum Algorithms, Quantum Optimization, Noisy Intermediate Scale Quantum Systems ( NISQ) 1 INTRODUCTION In the present time, the progress in developing quantum computers is very impressive. Many organizations are claiming their stacks in this space [1][2][3][4]. In today's world, the available quantum computers are at very early stages and not capable of handling complex quantum artificial intelligence/machine learning (qAI/qML) tasks [5]. But we still can harness their properties to run some of our quantum AI/ML algorithms more efficiently. In this sense, we can use the “Noisy Intermediate Scale Quantum Systems” (NISQ) [6] to serve the purpose. We can run the less complex quantum subroutines of a big qAI/qML in these kinds of quantum computers and use the results in the main qAI/qML problem-solving pipeline. This way we create a classical-quantum hybrid problem-solving eco-system in AI/ML space. We further can optimize the quantum subroutines at the quantum circuit level using ZX-calculus [7]. The optimized quantum circuits are less prone to the noisy results as the NISQ has to handle a lesser number of quantum gates calculations as compared to the original unoptimized quantum circuit. In this paper, we address an interesting problem in natural language processing (NLP) know as POS tagging [8] in a classical- quantum hybrid AI eco-system. Parts of speech (POS) tagging [8] is a very important task in Natural Language Processing (NLP) [9]. POS tagging is the process of assigning one of the parts of speech to a given word. Parts of speech include nouns, verbs, adverbs, adjectives, pronouns, conjunction, and their sub-categories. For example, considering the English tag-set “Penn Treebank” at the University of Pennsylvania [10]: :  → :  :  → :  :  → :  : ℎ → : ,  (1) Where the POS tags “NN”, “JJR”, “VB” are described as “noun”, “adjective/comparative”, “verb” respectively. Note that some words can have more than one tag associated with it. For example, “Chair” can be “NN” or “VB” depending on the context. The POS tagger is a function