The Promise of Quantum Computing for Artificial Intelligence and Machine Learning

Abstract

Faced with exponentially growing complexities and data demands necessary to train the artificial intelligence (AI) and machine-learning (ML) models, the classical paradigms of computing are successfully being restricted by computational hardness and scale limit capacities (Biamonte et al.,2017). The fact that quantum computing was developed over the past decades means that quantum-provisioned algorithms have the potential of exponential speeds improvement on some of the ML applications such as the ones involving optimization and sampling (Harrow et al., 2009). The paper looks into the possibility of quantum computing in AI and ML and discovering how quantum algorithms (Quantum Support Vector Machines (QSVM) and Variational Quantum Circuits) enhance the available ones (Schuld & Petruccione, 2018). Important hybrid quantum classical frameworks are revisited and reproduced with published, open databases of benchmarked metrics, such as accuracy and convergence rates (Otterbach et al., 2017). We can determine in our findings that quantum algorithms are theoretically and practically more resource-efficient and faster even though they are still very problematic to come up with some solutions in error correction, coherence, and use of qubits (Preskill, 2018). The study contributes to the already ongoing discourse concerning the application of quantum computational capabilities towards ameliorating the shortcomings of the current AI and identifies the directions in the scope of future investigations of hybrid and fully quantum machine learning systems (Cerezo et al.,2021). Quantum computing, artificial intelligence, machine learning, quantum algorithms, computational complexity, hybrid models, scalability